system analysis for the huntsville operational …
TRANSCRIPT
SYSTEM ANALYSISFOR THE
HUNTSVILLE OPERATIONAL SUPPORT CENTERDISTRIBUTED COMPUTER SYSTEM
ANNUAL REPORTMSSU-EIRS-EE-83-6
JUNE 1983 - JULY 1984
Submitted by:F. M. Ingels. Principal InvestigatorJ. Mauldin, Associate Investigator
Mississippi State UniversityElectrical Engineering Department
Mississippi State, MS 39762(601)325-3912
Submitted to:NASA/MSFC. Alabama
Technical Monitor: Frank Emmens, EB33(205)453-4629
NAS8-34906
July, 1984
iii
SYSTEM ANALYSIS
FOR THE
HUNTSVILLE OPERATIONAL SUPPORT CENTER
DISTRIBUTED COMPUTER SYSTEMS
SUMMARY
The Huntsvi l le Operations Support Center (HOSC) is a d is t r ibuted
computer system used to provide real time data acquisit ion, a n a l y s i s
and d i s p l a y during NASA space missions and to perform simulation and
study ac t iv i t ies during non-mission t i m e s . The p r i m a r y p u r p o s e of
the research is to provide a HOSC system s imula t ion model that may be
u s e d t o i n v e s t i g a t e t h e e f f e c t s - o f v a r i o u s H O S C s y s t e m
c o n f i g u r a t i o n s . Such a m o d e l w o u l d be v a l u a b l e in p l a n n i n g the
fu tu re growth of EOSC and in a sce r t a in ing the e f f e c t s of d a t a r a t e
variat ions, update table broadcas t ing and smart d isplay terminal da ta
requirements on the HOSC HYPERchannel network sys tem.
A s i m u l a t i o n model has been developed in PASCAL and resu l t s of
the s imula t ion model fo r v a r i o u s sy s t em c o n f i g u r a t i o n s h a v e been
o b t a i n e d . In th i s repor t a tu tor ia l of the model is p resen ted and
the results of simulation runs are p r e s e n t e d . Some ve ry h igh d a t a
r a t e s i t u a t i o n s w e r e s i m u l a t e d t o obse rve t h e e f f e c t s o f t h e
HYPERchannel switch over from contention to p r io r i ty mode under h i g h
channel loading.
Preceding Page Blank
l> -
iv
A computer tape was transported to NASA Huntsville and the
simulation program installed on the NASA computer. Appendix H is a
listing of the simulation code.
TABLE OF CONTENTS
SECTIONS
SUMMARY
TABLE OF CONTENTS
LIST OF TABLES
LIST OF FIGURES
LIST OF ACRONYMS
PAGE
iii
v
viii
ix
1. INTRODUCTION
1.1 HOSC System Overview
1.2 Research Objective
1
1
2
2. HOSC SYSTEM DESCRIPTION
2.1 Typical HOSC System Activities
2.2 HOSC System Components
4
5
9
3. HOSC MODEL DESCRIPTION
3.1 The Real System
3.2 Experimental Framework for the Base Model
3.3 Informal Description of the Base Model
41
41
41
43
4. SOFTWARE DESIGN DESCRIPTION
4.1 Scope
4.1.1 Major Software Functions
46
46
46
vi
TABLE OF CONTENTS (Continued)
SECTIONS PAGE
4.1.2 System Files 47
4.1.3 Major Design Constraints 48
4.2 Design Specification 48
4.2.1 Derived Software Structure 48
4.2.2 Interfaces within Structure 53
4.3 Module Description 54
4.3.1 ModSc 54
4.3.2 Initialize 59
4.3.3 CharacterizeNetwork 62
4.3.4 PrintNetDescription 66
4.3.5 FindNext 66
4.3.6 Picka 68
4.3.7 ACollision 70
4.3.3 DpdateClocks 71
4.3.9 PrintNetworkStats 76
4.3.10 ActivitySummary 78
For each module:
4.3.X.I Processing Narrative
4.3.X.2 Interface Description
4.3.X.3 Design Description
4.3.X.4 Modules Used
4.4 Validation Criteria 79
vii
TABLE OF CONTENTS (Continued)
SECTION PAGE
5. HOSC SYSTEM ANALYSIS AND CONCLUSIONS 80
5.1 Performance Bounds 80
5.2 Correlatiion with Simulation Results 85
6. REFERENCES AND BIBLIOGRAPHY 92
APPENDIX I SOFTWARE SOURCE LISTING 95
APPENDIX II SUMMARY OF ISO-OSI
MODEL OF DISTRIBUTED NETWORK 96
LIST OF TABLES
viii
PAGE
2.1 HOSC Data Transfers
2.2 Summary of DEC VAX 11/780 I/) Characteristics
2.3 Summary of PE 3240 I/O Characteristics
2.4 HYPERchannel Protocol Frames
4.1 Interactive Queries Posed by Initialize
4.2 Prompts for Network Characterization
5.1 Theoretical Upper Bounds on HYPERchannel Trunk
5.2 Cumulative Probability of Transmission MatrixDevice to Device
5.3 Configuration Parameters For the MultipleSimulation Runs
5.4 Statistical Results For The MultipleSimulation Runs
5.5 Comparisons of Theoretical Parameters S and UVersus Measured Results
7
17
20
37
60
63
84
87
38
89
91
ix
LIST OF FIGURES
FIGURES PAGE
2.1 Proposed HYPERchannel Network Configuration 6
2.2 Block Diagram of VAX 11/780 Computer 12
2.3 Basic Bus Configuration of VAX 11/780 13
2.4 Block Diagram of PE 3244 Computer 19
2.5 Multiple Trunking System 22
2.6 Delay Time Events 26
2.7 HYPERchannel Adapter Block Diagram 29
2.8 HYPERchannel Frame Formats 31
2.9 Message Only Frame Sequence 32
2.10 Message with Data Frame Sequence 33
4.1 Full Scale Model 49
4.2 First Level Refinement 49
4.3 Second Level Refinement 51
4.4 Third Level Refinement 52
5.1 Multiple Simulation Run Configuration 86
LIST OF ACRONYMS
ASCI
HDPBRCATVCPUCSMADDPDECDMAECIOETFEPI/OISOHSLNHOSCJSCKSCLPSMbpsMECOMERMPSMSFCNASANPRNSCODOSIPEPOCCRSSSBISRBSSNESTSUBAUBUSjisec
American Standard Code for InformationInterchangeBuffered Data PathBus RequestCommunity Area TelevisionCentral Processing UnitCarrier Sense Multiple AccessDirect Data PathDigital Equipment CorporationDirect Memory AccessExperiment Computer Input/OutputExternal TanksFront End ProcessorInput/OutputInternational Standards OrganizationHigh Speed Local NetworkHuntsville Operation Support CenterJohnson Space CenterKennedy Space CenterLaunch Processing SystemMegabits per secondMain Engine Cut OffMission Evaluation RoomMain Propulsion SystemMarshall Space Flight CenterNational Aeronautics and Space AdministrationNon-Processor RequestNetwork Systems CorporationOperational DataOpen Systems InterconnectionPerkin ElmerPayload Operations Control CenterRange Safety SystemSynchronous Backplane InterfaceSolid Rocket BoosterSpace Shuttle Main EngineShuttle Transport SystemUNIBUS AdapterUNIBUSMicrosecond
SECTION 1
Introduction
1.1 HOSC System Overview
Marsha l l Space Fl ight Center (MSFC), Huntsvil le, Alabama, has
i m p l e m e n t e d the Hun t sv i l l e Operat ions Support C e n t e r ( H O S C ) to
prov ide rea l - t ime acquisition, analysis, and display of data during
National Aeronautics and Space Administration (NASA) space m i s s i o n s .
The HOSC is a d i s t r i b u t e d c o m p u t e r s y s t e m composed of la rge
minicomputers and various peripheral equipment. P r i m a r i l y d e s i g n e d
to provide support for the Space Shuttle, Space Telescope, and Space
Labora to ry mi s s ions , UOSC has the inhe ren t f l e x i b i l i t y to be
e x p a n d e d and reconf igured to meet the needs of future missions, and
also to provide MSFC with a large computer resource that can be used
to support non-mission act ivi t ies .
During mi s s ions , the n e t w o r k c o m p u t e r s a re s e m i d e d i c a t e d to
p e r f o r m i n g s p e c i f i c m i s s i o n t a s k s , for e x a m p l e , one processor i s
responsible for the space shuttle main engine d a t a ana lys i s w h i l e a
s e p a r a t e p rocessor p rov ides backup capabi l i t ies for the same data .
The data arrives via satell i te communication l inks or d i r e c t ground
links f rom the Kennedy Space C e n t e r Fir ing Room at NASA's Kennedy
Space Center (KSC) in Florida or f r o m NASA' s Johnson Space C e n t e r
( J S C ) in T e x a s . D u r i n g a m i s s i o n , the c o m p u t e r s p r o v i d e da ta
analysis and presentation services for HOSC m i s s i o n support t e a m s .
In add i t ion to the m i s s i o n r e s p o n s i b i l i t i e s , HOSC also p rov ides
resource support for rou t ine non-mission a c t i v i t i e s such as the
P a y l o a d Opera t ions Cont ro l Cen te r (POCC) p r ep l ann ing a c t i v i t y .
Future non-mission activit ies may include computer r e source suppor t
fo r s u c h d i v e r s e t a s k s a s t he D i g i t a l Equipment C o r p o r a t i o n ' s
interactive graphics data base , IGDS, and the XEROX C o r p o r a t i o n ' s
SIGMA text-processing operation.
1.2 Research Objective
The diversi ty of tasks performed during mission and non-mission
operations at HOSC should be easy to p e r c e i v e since the f a c i l i t y
r e q u i r e s i n f o r m a t i o n - s h a r i n g b e t w e e n m u l t i p l e , i n d e p e n d e n t
processors. Based on the requirements of each mission support t e a m ,
the da ta t r a f f i c on the HOSC c o m p u t e r n e t w o r k may be composed of
short bursts of highrate data t rans fe rs or of s u s t a i n e d pe r iods of
high-rate da ta t r a n s f e r s . Furthermore, HOSC is a dynamic fac i l i ty
that must posssess the ability to adapt and change its ha rdware and/
software resources as particular mission requirements dictate.
In order to document the f lex ib i l i ty and e f f ic iency of HOSC, an
ana lys i s of the capabi l i t i es of the existing computer f ac i l i ty was
needed. Ideally, the analysis would not only o f f e r d o c u m e n t a t i o n of
exis t ing resources , but would also provide insight into fu ture HOSC
operat ions by p r e d i c t i n g p e r f o r m a n c e c h a r a c t e r i s t i c s in v a r i o u s
hypo the t i ca l mission scenarios. With this in mind, the primary goal
of the research e f fo r t was the development of a s i m u l a t i o n mode l of
the HOSC ne twork f a c i l i t y that could be manipulated to predict and
characterize the behavior of HOSC in a variety of situations.
The f i r s t s t ep in this analysis was the def ini t ion of a set of
resource configurations and mission scenar ios tha t migh t r e p r e s e n t
p l a u s i b l e g rowth s t ages of HOSC. One of these configurat ions is
outlined in Section 2 of this document. Included in the o u t l i n e are
the a v a i l a b l e and p r o j e c t e d ha rdware components along with their
operating c h a r a c t e r i s t i c s and f u n c t i o n s . The second s tep of the
ana lys i s , detailed in Section 3 and 4, involved the development of a
system model of HOSC and a s o f t w a r e s imula t ion of th is mode l . In
this s o f t w a r e s imulat ion, system parameters were varied to simulate
t h e c h a n g e s i n t h e o p e r a t i n g e n v i r o n m e n t a n d t h e s c o p e o f
r e s p o n s i b i l i t i e s of the HOSC. Finally, the results and conclusions
of these simulations were compiled and analyzed to provide a summary
of c u r r e n t ope ra t i ng c h a r a c t e r i s t i c s , as well as of the pro jec ted
operating characterist ics, of HOSC.
SECTION 2
HOSC System Description
As implied in the Int roduct ion, a p r imary task of HOSC is to
provide NASA eng ineers at MSFC wi th a near real-time data base of
critical informat ion daring m i s s i o n s of the Space Shut t le . This
information provides MSFC engineers and contractor personnel the data
they need as they act in a s u p p o r t c a p a c i t y to o t h e r m i s s i o n
specialists at other mission control sites such as KSC and ISC. HOSC
provides pre-launch support for the Space Shuttle Main Engine (SSME),
the Externa l Tanks (ET), the Solid Rocket Boosters (SRBs), the Main
Propuls ion System ( M P S ) , a n d t h e R a n g e S a f e t y S y s t e m ( R S S ) .
Concurrent w i t h mi s s ion r e s p o n s i b i l i t i e s , OOSC provides resource
support for mission experiments designed by MSFC specialists.
NASA has established the Flight Operations Support functions of
HOSC as follows:
Dur ing powered f l i g h t , the HOSC wi l l receiveonly data which is in the LPS (Launch P r o c e s s i n gSystem) at KSC. The Shuttle support team will be inthe HOSC during this phase of the mission and will bethe point of contact with the JSC Mission EvaluationRoom (HER) for problem discussion and r e so lu t ion asr e q u i r e d a n d w i l l b e o n ca l l d u r i n g o r b i t a loperations when applicable.
Fo l lowing comple t i on of the ac t i ve Shut t levehic le support a c t i v i t i e s , da ta i s r e c a l l e d asrequ i red for more de t a i l ed analys is , and in i t i a lp r e p a r a t i o n i s m a d e t o p r o v i d e s u p p o r t t opostflight evaluation [NASA82J.
HOSC is located in the west end of A-wing. Building 4663 at
MSFC. Figure 2.1 shows the functional components of the HOSC system
at the time of this writing.
2.1 Typical HOSC System Activities
Activities performed at HOSC can be grouped into two broad
categories, routine daily activities and miss ion/launch activties.
Table 2.1 provides a summary of the activities by category and also
provides the details of some typical data transfers. In addition to
the activites described below and in the table are the experimental
activities directed by the HOSC mission specialists. Since they
change from mission to mission, the activities are not included in
the following descriptions but must be characterized and considered
in the specific mission scenarios required for the total system
analysis.
POCC simulation is an ongoing simulation activity that utilizes
the computing resources of HOSC. This activity is not associated
directly with the real-time responsibilities of the network and is
thus considered to be a routine daily activity.
The impact of POCC simulation of the overall performance of the
network stems from the requirement of continuous data transfers
between network computers. This data is currently being transferred
between the MPS Primary Computer (VAX4) and the MPS Backup Computer
(VAX1). During each twenty four hour period, this activity requires
the transfer of 150,000 512 Byte blocks of data. This data is
transferred in two components; six times each day, an 8344-Byte block
o, •!->
OCJ
O
<u
CO Oii u-i
>-
T3
tC
Table 2.1. HOSC Data Transfers.
ROUTINE DAILY ACTIVITIESA. POCC Activity
Resources Involved:
Quantity of Data:
MPS Primary (VAX4).MPS Backup (VAX1).
150,000 512-Byte blocks in twocomponents. 6 8344-Byte blocks
and 100,000 512-Byte blocks.
B. ECIO Data Stream (Generated by POCC)Resources Involved: MIPS Backup (VAX1).
Spacelab 8/32.
Quantity of Data: 51.2 kilobyte per secondstream concurrent with POCC,
C. IGDS/SIGMA Activities (Proposed)Resources Involved: To be determined.
Quantity of Data: Unknown.
II. ROUTINE LAUNCH ACTIVITIESA. Routine Daily Activities (See Item I.)
B. Main Engine DataResources Involved:
Quantity of Data:
STS primary (PE 3244).MPS Backup (VAX1).
50 kilobit per secondstream (launch, minuys 8hours until MECO).
OD Data StreamResources Involved:
Quantity of Data:
FEP SSME (PE 3244).CSO Computer.STS Primary (PE 3244).MPS Backup (VAX1).
1923 kilobit per secondstream into FEP and then toCSO with transfer of 40 per-cent to STS Primary and MPSBackup (launch minuys 9 seesuntil MECO).
Engineering Display ChangesResources Involved: STS Primary (PE 3244).
STS Backup (PE 8/32).Spacelab 8/32 (PE 8/32)
Quantity of Data: Insignificant.
8
is t r ans fe r r ed (50.000 Bytes cumulat ive) , with another 100.000 512-
Byte blocks transmitted randomly but d i s t r i b u t e d evenly t h r o u g h o u t
the day.
POCC also generates a continuous 51.2 k i lob i t per second d a t a
s t r e a m known as the Exper iment Computer I npu t /Ou tpu t (ECIO) data
s t r e a m . This d a t a s t r eam i s ongoing and c o n c u r r e n t w i t h POCC
a c t i v i t y . ECIO da ta is t r a n s f e r r e d f r o m the MPS Backup Computer
(VAX1) to the Spacelab Computer (PE 8/32c).
The routine components of the launch day data activities are the
Main Engine Data Stream, the Operational D a t a (00) St ream, and the
E n g i n e e r i n g D i s p l a y Change r eques t s . The Main Engine D a t a i s
collected and d i s semina ted on launch day on ly . D a t a is funne led
t h r o u g h the ne twork to the MPS backup Computer (VAX1) f rom the
Shuttle Transport System (STS) Computer (PE 3244). From e igh t hours
b e f o r e launch until about twelve minutes a f t e r launch at Main Engine
Cut Off (MECO) . STS Pr imary accep t s a con t inuous 50 k i l ob i t per
second d a t a s t r eam d i r e c t l y f rom the KSC Firing Room and transfers
about 24 percent (12 kilobits per second) of the tota l s t r e am to MPS
Backup.
The OD Stream is a 192 kilobit per second d a t a s t r e a m a r r i v i n g
at the Front End Processor Space Shu t t l e M a i n Engine (FEP SSME)
Computer (PE 3244) on launch day concurrently w i t h some of the M a i n
Engine Da ta ( l aunch minus 9 seconds until MECO). This data arrives
from Goddard Space Flight Center, Maryland and is t r a n s f e r r e d over
the network to the CSO faci l i ty .
T h e E n g i n e e r i n g D i s p l a y C h a n g e a c t i v i t y i s a n a l m o s t
ins ign i f ican t add i t ion to the total network t ra f f ic . This activity
involves a transfer froa STS Primary to STS Backup and the Spacelab
8/32. This transfer consists of the name of each engineering console
foraat in the support faci l i ty that is changed daring this pre laanch
and launch period (launch minus 9 seconds to MECO).
2.2 HOSC System Components
HOSC may be thought of as simply a communications channel (Local
Area N e t w o r k ) for connec t ion of computer r e sources . The b a s i c
componen ts of the sys tem are the computers and the communications
processors that receive the transmitted data from telephone lines and
m i c r o w a v e l inks and the computer networking hardware that provides
the interconnection between the computers. The HOSC network hardware
is a h i g h s p e e d , loca l n e t w o r k m a n u f a c t u r e d by Ne twork System
Corpo ra t i on (NSC) ca l led HYPERchannel . H Y P E R c h a n n e l p r o v i d e s
c o m m u n i c a t i o n s s e r v i c e s b e t w e e n c o m p u t e r s f r o m a v a r i e t y o f
manufacturers such as Digital Equipment C o r p o r a t i o n ( D E C ) , Perk in-
Elmer ( P E ) , and UNIVAC. In order to provide a clearer understanding
of the resources involved at HOSC, a description of H Y P E R c h a n n e l is
given, along with a cursory look at two of the network computers that
provide the bulk of the processing power for the n e t w o r k : the DEC
VAX and the PE 3240. These mach ines are examined for sake of
information only, since from the point of v i e w of the subsequent
ana lys i s all ne twork users are considered to be nothing more than
data sources with fixed data generating a t t r ibutes . An in-depth look
10
is t h e n given to HTPERchannel ne twork since the success of the
simulat ion model w i l l r e s t d i r e c t l y on a c l e a r and a c c u r a t e
understanding of it.
The Digi ta l Equipment Corpora t ion 's VAX ser ies of compu te r s
s u p p o r t s a 32-bit word a rch i t ec tu re that e s t ab l i shes a v i r t u a l
address space of 4.3 b i l l ion bytes of user -addressable m e m o r y .
Throughput of data in the system is optimized by using a 32-bit high
speed data structure that ties together the central processor , main
memory, UNIBUS subsystem, MASSBUS subsystem and the DR780 high speed
direct memory access subsystem. This high speed da t a s t ruc tu re is
known as the Synchronous Backplane I n t e r f a c e (SBI) and a device
linked to it is known as a NEXUS. Each NEXDS rece ives eve ry SBI
t r a n s f e r ; electronic logic in the NEXUS determines whether it is the
designated receiver for the ongoing t r a n s f e r . D a t a t r a n s f e r s can
occur from CPU to memory, from I/O controller to memory subsystem, or
from CPU to I/O controller, with maximum aggregate transfer ra tes of
13.3 megaby te s per second. These transfer rates are limited by the
following factors:
200 nanoseconds/cycle = 5 million cycles per second
each cycle can carry an entire byte of data or
address representing a memory request
. one cycle is used to request eight bytes of data
to be read or written, and two cycles are used to
carry data at four bytes per cycle
5 million cycles/second * 4 bytes/cycle = 20 million
bytes/second
11
20 * 2/3 (1 of every 3 cycles is an address) = 13.3
million bytes per second [DEC80]
As is shown in Figure 2.2. a VAX computer may interface with
more than one memory subsystem. In the case of a two-controller,
interleaved memory configuration, the computer would also have two
NEXUS memory controllers on the SBI.
The DNIBUS Subsystem (UBUS) is a high speed, asynchronous data
system that allows communication between peripheral hardware and the
VAX computer. The VAX 11/780 is capable of supporting four UBUSes:
one is standard, but three more are optional (Figure 2.3). The UBUS
is connected to the SBI through a UBUS Adapter (DBA) which performs
several important services that are transparent to the user. The UBA
provides:
access to UBUS address space from the SBI
mapping of the UBUS addresses to SBI addresses/
for the direct memory access (DMA) transfers to
the system memory
data transfer paths for UNIBUS device access to
random SBI memory addresses and high speed transfer
for devices that transfer to consecutive, increasing
addresses
UNIBUS interrupt fielding
UNIBUS priority arbitration
12
, ,. ^^^
CJ
(Xfa
Ha:O >a. o
•4
£M OH 2»J bP 2
J_•
OSO >P-: aiM OH 2hJ UjP X
ar.
>- i-Cri _
*•i-1
3
IICCJ
OSWH-jT
OS 0
9 ?iit r~"
s: q
o
r-4 J3iH P.;
1 U
J5
CO W^ f— W o
CO 25>* CCO CJ
XIo S
^C E^m -3<N
^£
S^^CM
srvp CO-3" E-1 <>•!N
-C
^
CN
IT!
COh- '
W COH 0C ZS 0
OS MP
X
s.
!
n*'CJ
COctf£-J
I-Ji—30erfH
OCJ
gSs-CN
MCO :
H U(X tu< CO
£
C
1— 1
(__j
••H
O
i— i
X
€p^\Q.
^O
f1QZL\
I
\
sna S3
I
rsaaI^
iU
IJJ CO CH f-1 C< crj- fTT ^W H ,
ga l 5
< <•/•
PuO
oz
3a.EOuCoo
U-lo
ocCO
uo
oC•Hfe.
13
Qa esCL Oc 3
h-i 33
-sf
C6
cd p.,CLi '
uCJ EdEd to
Cd CO•z. oH-l O Z
<Cd ZJ OCJ O O>* O <TCJ CO r-l
<cZ
k-KL_...__l
cuCJ
/I—
M Cd 'CO
Cd ~~-
< £OCd mOS •O mO rH
b^ r <
CdU<feasCdHZl__l
2:<2J
p.CJ
oaCO3OZ0OS3:CJzCO
^^•^ fc
^
OS. 0 Ed
r^. p ,
Q a• <c
L . .
CO3 2S
« » C£ L^I
co a.< a
: z <[
DCc/: CEJ
• » — » c_32 Q_
J h^ <!
s <
CO
^^
Zc ^H
0
. .. i
CO0 Cd« CJ
Q CdO
XC;^H
. |
- y-t
O
Co
•rH*Jca
" 1 r* ft
S 1, - . ! " *
!j
CO3 C/1CC Cd
CJCO '— '
< Eds: c
3
•nU-j
OU
CO3
esX— s
^ U
W < J2 X^ CQ s: o-Q S LO 2C
OS XO M <: r^CN D- S •
rsl
CON_^
rH »^
COCO Cd3 CJ
3 a
w2oc•H
CO
n »•2 ^
i— i
14
The address-mapping function is especially important because the
U B U S h a s o n l y e i g h t e e n d a t a l i n e s p r o v i d i n g a n a p p a r e n t
memory-addressing capab i l i ty of only 200 k i lobytes . Through its
address mapping abilities, the DBA provides the capabili ty of mapping
the UNIBUS addresses into the SBI addresses so that the f u l l sys tem
memory of 16 ar ray boards of 256 kilobytes each (4 gigabytes total)
can be accessed.
T h e U B A a c c e p t s t w o f o r m s o f i n p u t f r o m t h e U B U S :
hardware-generated in te r rup ts and d i rec t memory acces s reques t s .
Each device connected to the UBUS may use one of five priority levels
for requesting bus service. A non-processor request (NPR), which is
the highest priority request, may be used when the device requests an
operation that does not requi re p rocessor i n t e rven t ion , such as a
d i rec t memory access transfer to memory or to some other device. A
bus request (BR) i s thus g e n e r a t e d w h e n the d e v i c e w i s h e s to
interrupt the UBA for service. Such service might be a CPU-directed
data t r a n s f e r or a f l a g i n d i c a t i n g the e x i s t e n c e of an e r r o r
condi t ion at the per iphera l . Since there are only five priority
levels and more than one device may be connec t ed to a s p e c i f i c
request level , if more than one device makes the same request , the
device that is electrically closest to the UBA r e c e i v e s the h ighes t
priority.
The NPR request for d i rec t memory access is a very impor t an t
f e a t u r e of the UBUS subsystem. These DMA transfers can be divided
into two groups: random access of noncont iguous a d d r e s s e s and
sequent ia l access of sequentially increasing addresses. For random
15
access, each UBDS transfer is made through the Direct Data Path (DDP,
one per UBUS) and is mapped into an SBI transfer. This procedure
allows only one word of data to be transferred during a single SBI
cycle. For devices capable of requesting sequential access services,
use is made of the Buffered Data Path (BDP). Each UNIBDS provides
fifteen such BDPs which store the data so that four UBUS transfers
are performed for each SBI transfer.
The DDP must be used by devices not transferring to
consecutively increasing addresses or by devices that mix read and
write functions. The maximum throughput via the DDP is about 425
kilowords per second for write operations and about 316 kilowords per
second for each read operation. These rates will decrease as other
SBI activity increases [DEC80].
Maximum published throughput via the BDP is about 695 kilowords
per seconds for both the read and write operations, but realistic
throughput rates of only 1.5 megabits per second are actually
expected with this figure degrading as other SBI activity increases
[WILK82.MCMU82]. BDP transfers are restricted to block transfers
with blocks of length greater than one byte. All transfers within
the block must be to consecutive and increasing addresses and all
transfers must be of the same function, either read or write.
The MASSBUS subsystem and the DR780 high performance, 32-bit
parallel interface will not be described in this summary, since an
understanding of their functional characteristics is not needed to
determine their relative impacts on the HYPERchannel network. The
influence of both may be felt indirectly, however, since activity on
16
the MASSBDS or DR780 will translate to SBI activity which will a f fec t
DOP and BDP transfers rates as previously described.
The VAX CPU will also not be desc r ibed in de t a i l but severa l
c o m m e n t s may be m a d e a b o u t the CPU and i ts e f f e c t s on sys tem
throughput. The CPU represents the most intensive t r a f f i c load on
the memory s u b s y s t e m and hence on the SBI. Obviously, if the
processor is engaged in in tens ive comput ing , it wi l l reques t data
much more o f t e n than it wi l l wri te data , and this memory access
represents substantial SBI ac t iv i ty . For tuna te ly , the large cache
memory of eight kilobytes available to the CPU is able to reduce the
SBI t ra f f ic load considerably. In addition to the ef fec ts of the CPU
of SBI ac t iv i ty , the SBI t r a f f i c from other sources may also a f f e c t
the CPU eff iciency. Published f igures indicate that in a system with
two memory control lers the processor wi l l be slowed by about four
percent per averaged megabyte per second of I/O t r a f f i c , whi le the
impact of a s ingle memory controller is to slow the processor by a
f a c t o r varying f rom two to four . Table 2 .2 s u m m a r i z e s the I /O
characteristics of the DEC VAX 11/780 processing system.
TABLE 2.2 SUMMARY OF DEC VAX 11/780 I/O CHARACTERISTICS
17
PROCESSOR
MAIN MEMORY:
Virtual Address Space:
Cycle Times:
I/O DNIBUS ADAPTER
Maximum aggregate rate:
Buffered Data Paths (BDP):
Direct Data Path (DDP):
32-bit word architecture
4.3 billion bytes.
800 nanoseconds per 64 bitread.
1400 nanoseconds per 64 bitwrite.
1.5 Mbyte/sec through BP
15 total, 8 byte buffer ineach.
695 kilowords/sec read andwrite.
Used for fast data transfers.
425 kilowords/sec for write316 kilowords/sec for readUsed for transfers to noncon-secutive memory locations.
18
The Pe rk in E l m e r 3240 ser ies of c o m p u t e r s is a lso a h igh
throughput machine with a 32-bit a r ch i t ec tu re . The HOSC c u r r e n t l y
uses two PE3244 machines with primary responsibilities as front end
processors receiving real-time data streams from the KSC firing room.
The 3244 memory subsystems is organized into banks, each capable
of handl ing four megabytes of addressable memory. Total sys tem
m e m o r y r a n g e s f r o m 256 k i loby te s in one bank to a f u l l sys tem
complement of four 4-megabyte banks, for a maximum of 16 megabytes of
addressable memory . All memory is connected to a common memory bus
which consists of two unidirectional, asynchronous, 32-bit b u s s e s .
One bus is d e d i c a t e d to memory w r i t e f u n c t i o n s and the other Is
dedicated to memory read functions (Figure 2.4).
Inpu t /Outpu t is accomplished using five external communication
busses: one multiplexer bus for medium speed d e v i c e s and a m a x i m u m
of f o u r high-speed Di rec t Memory Access (DMA) busses tha t each
support e ight high speed b i d i r e c t i o n a l p o r t s . Each DMA por t is
con t ro l l ed by a se lec tor channel t ied to the mult iplexer bus that
controls and terminates transfers through the CPU. Once the channe l
is a c t i v a t e d , the processor is r e l eased and is f r e e to cont inue
processing on an unrelated task. Publ ished I/O t r a n s f e r r a t e s for
the PE3244 DMA bus indicate that transfer rates of up to 10 megabytes
per second in the burst mode are possible for each DMA bus [PEC81].
The I/O charac te r i s t i c s of the PE3240 series is summarized in Table
2.3.
19
16 MB MAX
BANK 1 BANK 4\
MEMORY ACCESS500 NANOSEC
COMMON MEMORY BUS
MEMORY ACCESSTRANSLATOR
SELECTOR
(CONSOLE CRT
QMULTI-CRTS
r CARDREADER
v:=>33
CC
W
PUh-(H
REGISTER STACKS(8 SETS OF 16x32 BITS)
LOOK AHEADSTACKS
(2 SETS OF2x32 BITS)CPU
Figure 2.4 Block Diagram of PE 3244 Computer
20
TABLE 2.3. SUMMARY OF PERKIN ELMER 3240 I/OCHARACTERISTICS
PROCESSOR: 32 bit/word.
MAIN MEMORY
Virtual Address Space: 16 Megabytes.
Cycle Time: 500 nanoseconds.
DMA BUS DATA TRANSFER RATE
Burst Mode: 10 megabytes per secondwith a maximum of 4 DMA
bussesper system.
21
The Network Systems Corporation (NSC) HYPERchannel high speed
local area computer network is a 50 Megabit per second (Mbps)
baseband communication system employing a Carrier Sense Multiple
Access (CSMA) contention scheme with prioritized, staggered delays.
Network messages called frame sequences are broadcast over a 75-ohm
CATV cable (called a trunk) at a fixed 50 Mbps rate, using a phase
modulation encoded Manchester code and received by network components
at various nondirectional taps or ports along the trunk (Figure 2.5).
Access to a trunk can be gained only at a port and is controlled by a
network adapter present at that port. NSC supplies three types of
network adapters: processor, device, and link. The processor
adapter is designed to allow trunk access under a strict set of
network protocol rules for intelligent devices such as minicomputers
and mainframe processors. NSC designates these adapters as A400
HYPERchannel Adapters which will hereinafter be referred to as A400
adapters, or more simply as adapters. The device adapter provides
trunk access for devices such as printers, dumb terminals, and mass
storage devices. Network to network communications are made possible
through the link adapters. Due to the nature of the communication
environment at HOSC, the only adapter included in this model is the
processor adapter [THOR83].
22
CD
ccr;
•H
c3
OJ
c.
0)
OC
23
When an adapter receives a request from a host processor for
trunk usage, the adapter assembles the host data along with the
appropriate control and destination information into a package and
attempts to transmit the first frame packet of the package on the
trunk. At the trunk level, adapters are capable of carrier sensing,
which means that before a trunk transmission is attempted the adapter
first determines whether or not a carrier is present on the trunk.
The existence of a carrier indicates an ongoing transmission and the
ready adapter defers its request in favor of the ongoing
communication. Carrier sensing, however, in itself is not sufficient
to overcome the problems that arise when two or more adapters attempt
simultaneous or near simultaneous trunk transmissions. This
condition, which usually results in unintelligible communications and
subsequent loss of transmitted frames, is known as a collision. To
overcome this shortcoming and increase the 'hospitality' of the
network, the adapters provide several other services that are
adequately summarized in the following list from Donnelly and Wey.
... [The] ports supplied by NSC are packaged in anadapter that includes some high-speed buffer memory anda microprocessor that:
Decouples the 50-Mb it/sec trunk speed fromthe possibly slower speeds of the inter-faces to the attached systems.Handles duplicate detection on the trunk.Supplies a flow of control mechanism.Recovers gracefully from various errorconditions.Allows flexibility in communicating withattached systems [DONN79].
24
As mentioned previously, carrier sensing in itself provides
inadequate contingency processing actions for adapters sensing a busy
trunk. By using timers in each adapter, HTPERchannel is able to
supplement the carrier sensing by providing linear priority ordering
of adapters on each trunk. Upon detection of a busy trunk, an
adapter goes into a waiting state by setting a series of delay clocks
that will determine the time of the next possible transmission
attempt. This action occurs in every adapter any time a carrier is
sensed on the trunk regardless of whether the adapter is actually
ready to transmit. As soon as the carrier drops, indicating that a
current frame transmission has passed and that the trunk is free, the
clock countdown toward the next transmission attempt is initiated.
If at anytime during the delay state, the trunk is again sensed busy,
the delay times are reset and reinitiated as soon as the carrier
drops.
As the traffic of transmitted packets increases toward a level
where the trunk is in a state of constant demand, this series of
delays will effectively slot the transmission medium, thereby
allowing prioritized time slots in which each adapter may gain trunk
access. The series of delays is called the total delay and may be
composed of three separate delay events: fixed delay, priority
delay, and end delay. The fixed delay represents the maximum amount
of time required to transmit a frame between the two adapters
separated by the longest physical length of cable. In addition to
being a component of the total delay event, the fixed delay also
represents the maximum amount of time an adapter must wait to receive
25
a response frame from any other adapter on the trunk and, as such,
prevents additional adapters from t r a n s m i t t i n g un t i l the r e c e i v i n g
adapter has had t ime to send a response frame to the transmitting
adapter. This delay in microseconds ( u s e c ) is d e t e r m i n e d for each
trunk in the network by the following equation [NSC79]:
.,. . n , Total Trunk Length ( f t )Fixed Delay = TT * * ' + 13(usec) *U
This value is set into switches in the adapter and is recommended by
NSC to be the same for each adapter on a particular trunk.
The priority delay (DLy) allows the linear priority ordering of
adapters on the trunk. This delay will permit adap te r s w i t h h igher
t r u n k a c c e s s i o n p r i o r i t i e s p r o p o r t i o n a l l y more trunk t ime than
a d a p t e r s w i t h l o w e r p r i o r i t i e s . T h i s p a r a m e t e r i s a n
i n s t a l l a t i o n - e s t a b l i s h e d value that can be d i f f e r e n t for each
adapter. Computation of the priori ty de l ays b e g i n by choos ing the
a d a p t e r ( s ) to have zero p r io r i ty de lay and then es tabl i sh ing the
p r i o r i t y c lasses f o r t h e r e m a i n i n g a d a p t e r s a c c o r d i n g t o t h e
following equation [NSC79]:
(priority delay (cable length ( f t ) to
Priority Dly - 10 + °J »"*"•*"* + *"* hi^est a dP t r>(usec) adapter) 40
Once the priority delay has experienced, the adapter may capture the
trunk. At this level, the trunk is in a pure p r i o r i t y scheme w i t h
each adap te r a l lowed a time slot for transmission (Figure 2 .6) . At
26
zo
&JH CS CO H-iU OS
ua P..
j Hw zQ W
c wz
I
OSp-
cc
HH- >-OS <O "
OScOSp-
i-i _ M [I]
^ OS Cp^
fcccwX
CrtHOS
25>H en
Q• wXM Q
zos
PS HUH W0. X< aQ K< C/l
aoI—Ias
<w
OSp-
c01
OJc•HH
OS!Xjh-1OS
3u
asUJH-(OS ZOS O
0)i-l3
27
the conclusion of the entire priority delay period, all adap te r s are
free to contend for the trunk.
If an adapter completes a transmission during its priori ty delay
t ime , it may conceivably at tempt to regain the trunk sometime during
the remainder o f t he p r i o r i t y d e l a y p e r i o d . To p r e v e n t t h i s
occur rence , each adapter is provided with a wait flip-flop, which is
enabled by a wire jumper on the printed circuit board assembly in the
adap te r . If the user desires to allow each adapter only one actual
access to the trunk during any one priori ty delay period, this jumper
may be removed to enable the wait flip-flop. This action forces the
adapter to cycle through an end delay before again contending for the
trunk thus a l lowing the adapters in the other priority classes an
opportunity to gain access to the trunk. Therefore, the end delay is
the t ime f o l l o w i n g the f i x e d de lay and the priori ty delay in which
all adapters on the trunk have free access to the trunk. This d e l a y
is c a l c u l a t e d for each adap te r as shown in the following equation
(NSC79]:
(cable to length ( f t ) from(priority dly adptr to far thes t adptr)
End Dly = 10 + of lowest + ftadptr) 4U
The timing control just described is i m p l e m e n t e d and a u g m e n t e d
by a r e s i d e n t m i c r o p r o c e s s o r in each a d a p t e r . In a d d i t i o n to
protocol management , these mic roprocesso r s p e r f o r m severa l o ther
c r i t i ca l func t ions . The adap te r microprocessor receives parallel
data from the host processor r eques t i ng t runk acces s , b u f f e r s the
28
data , forms it into network frame sequences, and then transmits the
packe ts ser ia l ly over the trunk under the p rev ious ly d e s c r i b e d
pro toco l . Acting as receiver, the adapter collects the serial data
from the trunk and a s sembles the d a t a in to pa ra l l e l p a c k e t s for
t r a n s f e r to the a p p r o p r i a t e host p r o c e s s o r . The f u n c t i o n a l
characteristics of the adapter are shown in Figure 2.7.
The e f f e c t of the desc r ibed p r io r i ty pro tocol is to lessen
considerably the possibility that an adapter will attempt to b e g i n a
c o m m u n i c a t i o n s e s s i o n when the t runk is engaged in an ongoing
transmission. Although the time window for such a collision has been
closed to a mere c r ack , a smal l t ime slot still exists in which a
collision may occur. This occasion most often arises when two or more
adap t e r s become ready to transmit almost simultaneously, that is to
say they seek to access the trunk within a time interval that is less
than the t r u n k end- to -end t r a n s m i s s i o n t ime. This r e s u l t a n t
col l i s ion and loss of da ta is thus a very real p o s s i b i l i t y and
contingent actions are needed.
HYPERchannel provides a r e t r y m e c h a n i s m to r ecove r f r o m such
co l l i s ions . Under this scheme, an adapter is able to detect when a
col l is ion has occurred, and the adap te r r e s c h e d u l e s s u b s e q u e n t
t ransmiss ion a t t e m p t s . In order to understand more completely this
re t ry mechanism, a more t h o r o u g h u n d e r s t a n d i n g o f the p a c k e t
transmission apparatus is needed.
The f i r s t s t e p i n c o m m u n i c a t i n g o v e r t he t r u n k i s t he
e s t a b l i s h m e n t o f a v i r t u a l c i r c u i t b e t w e e n the two or more
communicating adapters. When an adapter rece ives a request f rom a
29
tJ atcj: fai
as < H — ! >-U OS CO C OS
CS H O i"S ha ES Ou: c. E- z cj E-i sE- < cr. 3 a: z Mp. a os PS c r:< < H cjG
en en J3 co Op^ u; cs»-: as E-r ; ^ JT ___3 GOZ < cj
i
H en zCA! -3 O OSCS CO. ha ul
cj en Hc* r~. , . z enIs, OS ^ ia har- P- H CJ5- X t-'
< CS
C Z * .C O W \
5 U2 \O en ^ha oT ?
|i. * i* t f
?*-i
wE-3
ha p. E-iZ Z enh- o O"Z. CJ ~£
o—o-
>a < CS OS ^Nry.os cs z t: oa c o fa. rfa. S: h- fa. w rv__fa. M E-; 3 sr v*^3 S 0- =0aa o
3
H
Q
O< OS _JS E- O -Q Z OS
C HCJ Z fa]
+ CJ ha1 cn
/ ' ^
t ^
\\*^~
^^
L/
D^"
SSSJ
y INT
ERFA
CES
2 C
ON
TR
OL
LIN
ES
NC
LU
DIM
G
16
INP
UT
&
:~:J.
cn
i OSz uCS Z
zh-l
COM a-;z as a— • W HhJ E- P-
P- <-I < QC 0 < MOS < CJE- co <Z W 3 t,
** O U fa; esCJ ha ^ t:
> O Hrg W 3 Z
6
OU
TP
UT
LIN
ES
5
75
fi
TW
IST
ED
P
AIR
S
,E t T
su
pp
ort
s
the
m
axi
mu
m
ate
tr
an
sfe
r ra
tes
.
cu4-1 laa o« CO-s in< cu
CJc oO 1-<r c.
cu en,= 0
•• £- £
XD
S
zCSH
C
0CJ
•H
O
^O
ac.ct
V4~
)-f
i.
C
r1"C1
J.
c7- .2
sTJ-~^
"Ja4_
1•• ccu ^3 1-Hrj sj> S3
CSM C^03O ti
a.
1 C cu--~ r-l
Ua 2
O «J
4-1 CJen•H <
C Q
o -a
•a•H en .en u paj 2 «
la CU -HCJ Cc* f*
c c ^
*^ QCU CO - JjCJ C C 5
•i-i ai o w
> 4-1 -r-l
a X cn w
Q tU la CD
. S a••' en c cfl
<• ~ a o -a3 cj ej . <
*_l CJ ^H CJ i
CJ - er r-| -r-i j_!
u c en cej cj C•i-i O >- en «)T3 -H CU CXI C -C
*j e a. a oo a S3 -s, u 3S
) 4J CU >a i— 1 4-1 CiJ1 r-i U-l CJ J2,
. en er. Ua la c -r< S O CU 3j o • C j<: ec u' "34-1 2S C X 4J
• .? x -• - ^ »7 ^] _ ^ 3 "cc ^ w "§ _c' £ ^3 i— i cn i o cj •> en u
" 0 " S, i, -H -H 3 Ji B^ C C — c j 4 j e o " 3 ^ i c j .
•H -H ci s c eo -~- a rTo a) 4J ^- -f-i cu C —
JO U O - ^ r J g , _ | «v , u
j K c u c u enG ^ -~en^ : ha) — - er, en a i—> c £ "3
i cj en en >» cu o i a c n orai4 j c u c u £ > - - ^ 4 - i c n
" *-I^ • 3 v - ^ " ' C ' r a " 3oi r-i ca > ra .en la
• vo scr u a c cu^ CN i o 13 en o S
l to c cj ocn en r-i o oc
U 3 O O v-4 CUr CM j-> cj la cn -o -ar r-> •-. O 4-1 'J) -r-i laj -^. c >a c -H en ai oo 3 a. o E 2
O o en . _<i: LJ cj u c c cj: •"** U! CJ "3 r3 3 CU
Pa C -r. C I- t- a=Z. R 4-i E-1 cj
C
30
host for trunk access, the adapter first reserves itself, thereby
marking itself as unable to become part of another communication
session. After reserving itself, the adapter attempts to reserve the
prospective receiving adapter(s) by sending a Set Reserve frame on
the trunk. If the receiving adapter is idle, it will respond by
setting its reserved flag and transmitting a ready-to-receive
response frame to the initiating adapter. (Figure 2.8 shows the
contents of the reservation frame sequences.) After the reservation
sequence, the initiating adapter releases the trunk (both adapters
remaining reserved but other adapters could now use the trunk) while
it forms the message frame sequences containing the data. These data
sequences are of two classes, message only sequences and message with
data sequences. Message only sequences allow the the transmission of
a message proper, which includes the frame sequence data, access
codes, destination codes, and a small amount of data up to a total
frame sequence length of 64 bytes (Figure 2.9). For larger data
transfers, a message with data sequences is utilized in which data
are transmitted in blocks as large as two kilobytes (kbytes) in
length (Figure 2.10). (Four kbytes are an option when the adapters
have increased buffer capacity). After the communicating adapters
have been reserved with a message proper sequence and trunk access
has been regained for the data transfer, a Copy Register frame is
sent to the receiving adapter and, if successfully responded to, the
sending adapter captures the trunk and sends the two kbyte data
block. Subsequent data blocks are sent until the entire message has
been sent. At this time, an End of Data message is sent by the
31
field
leading syncsframe typeaccess codeto addressfrom addressfunction codemessage lengthcheck wordtrailing syncs
length(bytes)
12 - 1812112228
field
leading syncsframe typeaccess codeto addressfrom addressfunction codemessage lengthcheck wordtrailing syncs
length
312112228
Transmission Frame Format Response Frame Format (No Data)
field length
leading syncs 12 - 18frame type 1access code 2to address 1from address 1function cde 2message length 2check word 2message/data 8-64/0-2Kmssg/datacheckword 2trailing syncs 8Message Proper and Associated
Data Frame Format
field
leading syncsframe typeaccess codeto addressfrom addressfunction codemessage lengthcheck wordregister information
length
3121122216
checkword 2trailing syncs 8Response Frame Format (with Data
Figure 2.8. Frame Formats.
32
SENDER
Set Reserve
Copy Registers
Message Proper-
RECEIVER
•Response
-Response (Inc. reg. Inform.)
Response
Clear Flag 8
-Response
•Frames enclosed within bracket are transmitted without interference
by beginning their transmission in the fixed delay period.
Figure 2.9. Message-only Frame Sequence
33
SENDER
Delay
( P r e p a r a t i o n to
Transmit)
Set Reserve
CHANNEL CHANNEL INTERVAL
RECEIVER HELD FREE RELEASED DURATION
Delay
Response
Copy Registers
Message
(64 Bytes)
Clear Flag 8
Response (Reg. Info.)
Delay
Response
Response
64
12
23
74.5
20
Figure 2.10. Timings for Message-only Sequences.
(Transmission times do not include 1.32
microsecond per foot propagation delay.)
34
0 ncuao1-1PL,
cu00mCO
64 Ss
X1
X
20472048
*
+
4000
I \
| ( <
Ass
ocia
ted
D
ata
Blo
ckV
N
L
\ \
\
->
^*»••
T
MM
^
y
1 Set Reserve i
1 Copy Registers •
V PAi'M'iirff Moccogo
/ (Send Message)
* Set Flag A »
Clear Flag 8. »
«. _ _Response M.Copy Registers—*
\ *V. Receive Data/ (Send Assoc. data)
| * P1*«r P]^g 0
^ • ;Response •Copy Registers—*
^ Receive Data rr (Send assoc. data)
* Clear Flag 9 »
Response »Clear Flag A m
*Responseb
Response Incl. Reg. Info.
Response
Response
ResponseClear Flag 9
Response (+ Register)
Response
ResponseClear Flag 9
Response (^Register)
Response
ResponseClear Flag 9
Response
•Frames enclosed within brackets are transmitted withoutinterference by beginning their transmission in the fixed delayperiod.
Figure 2.10 (Cent.). Message with Data Frame Sequence.
35
RECEIVER
INTERVALCHANNEL CHANNEL LENGTHS
ISED HELD (MICROSEC)Preparation to transmitSet Reserve »
Delayfnnv Ppff i *if'i*i* -^
Message t
Set Flag A .... *
Clear Flag 8 *Delay
Delay
As soCi Data __ ._. »(2K bytes)
Clear Flag 9 .. .. »
Delay
Pf»«nrmc«» -
Clfar Flag A .— -»
Delay
Response
Response
Response
ResponseResponse
Clear Flag 9
Response
Response
Response
— flear F1 ao Q
I•
•
for each associate
Response
64
12 —23
96.5
64
1229 ""
412
3412
• •
• *
d data frameI _
1220
Figure 2.10 (Cent.). Timings for Message-with-data sequences.(Transmission times do not include 2.32microsecond per foot propagation delay.)
sending adapter, and both adapters release themselves, returning to
an unreserved state. A brief summary of the HYPER channel protocol
frames is given in Table 2.4.
During the communication session just described, the trunk is
released after each data frame (two-kbyte packet) sequence.
Conseqently, the trunk must be regained for the transmission of each
two-kbyte data block. This allows other adapters a chance to contend
for the trunk during the time a message sequence is being formed.
Adapters also provide an alternative to this transmission mode known
as burst mode. Under this scheme, the trunk is not relenquished
between data block transfers but is retained until the entire message
has been transmitted. This mode will not be described in further
detail since it is not used in the HOSC network.
Each adapter is capable of interfacing with and controlling the
trunk transmission requests of from one to four minicomputer hosts of
the same or of differing manufacturers. This interface is effected
by an NSC Perepheral Interface board resident in each minicomputer
host, a Device Interface for each computer, located in the A400
adapter and the associated cabling and connectors between the adapter
and the host processor. Whenever an adapter is not actively engaged
in some function, the adapter microprocessor scans the device
interface ports in a round robin polling sequence for a request to
provide some network service. When a request is detected, the
scanning is suspended and the service is performed. After completing
the service task, the polling process is resumed.
37
TABLE 2.4HYPERchannel Protocol Figures
Transmission Frames
Set Reserve - Reserve receiving adapter
Copy Registers - Request for Register information
Clear Flag 8 - When set. Flag 8 indicates that theadapter is ready to receive a frame.When clear, it indicates that theframe has been received.
Set Flag A - Flag A when set indicates that theClear Flag A receiver will receive associated
data. It is cleared when the lastblock of associated data is received.
Clear Flag 9 - When sst in the receiver. Flag 9indicates that the receiver is readyto receive data. When set in thetransmitter, it indicates that thetransmitter is ready to send data.When cleared in receiver, itindicates that the data block hasbeen received. When cleared in thetransmitter, it indicates that datacan now be transmitted.
Data and Message Frames
Message Proper - Used to transmit host messages oflength less than or equal to 65bytes.
Associated Data - Used to transmit data blocks oflength up to 2E bytes.
Response Frames
Response - A receiving adapter sends a responseframe for each non-response framereceived. Its purpose is toacknowledge received frames or, as inthe case of the resonse to a copyregisters frame, to send requesteddata.
38
In addition to the four device ports, the adapter may access
from one to four different trunks through separate trunk interface
boards resident in the adapter (Refer to Figure 2.6). When a network
service requiring a trunk activity is required, the host processor
sends a request to the adapter containing a list of trunks to try.
The adapter microprocessor then polls viable trunks in search of an
open transmission path. This selection is also made from a round
robin polling of trunks connecting the transmitting adapter and the
prospective receiving adapter, with each poll takling about 5.5 fisec
[FRAN82].
If an adapter attempts the reservation of a disabled or
previously reserved adapter, a Reject Reservation frame is sent by
the reserved adapter. This rejection will cause the requesting
adapter to put itself into a reservation retry loop in which entries
are scheduled after a binary exponential delay. That is, after the
first rejection, the sender delays 1 usec before attempting anothert
reservation. If the requested adapter is still not free, the
requesting adapter will wait 2 (isec more before attempting the third
reservation request. This scheme is repeated, with each subsequent
retry delayed an amount of time equal to twice the previous delay
until the adapter is reserved or until the retry delay reaches 128
usec. At this point, theretry delay clock is reset to 1 usec and the
process repeated for a retry count number of times. The retry count
is an adapter specific value based on the six low order bits of the
manually set adapter unit number and is in the range of 55 to 250.
Thus, 64 unique retry counts are possible. If after retry-count
39
delay sequences the adapter is still not able to secure the desired
reservation, the adapter will return to the idle state, informing the
requesting host of the inability to transmit.
This retry mechanism is enhanced by the ability of the
requesting adapter to determine the identity of the receiving
adapter's reservation partner. If the requesting adapter discovers
that the reservation lock up is caused by both adapters attempting to
simultaneously reserve each other, the requesting adapter aborts its
reservation attempt and releases itself for reservation by the other
adapter.
Although the retry scheme provides some advantages, it also
leaves unresolved two important problems. First, due to the slowness
of the back-off mechanism (about SO usec are needed for an adapter to
realize the simultaneous reservation problem and to react), both
adapters may back off. Consequently, neither host is permitted to
transmit on the trunk.
The second and more serious problem presents itself when a
reservation request loop is formed. This occurs when three or more
adapters attempt to reserve each other simultaneously. For example.
Adapter A attempts to reserve Adapter B, who attempts to reserve
Adapter C, who in turn attempts to reserve Adapter A. The
exponential delay mechanism is initiated in each adapter and proceeds
to completion in each adapter, terminated only by the retry count.
Obviously, serious degradation of the trunk throughout may be
predicted during these reservation request loop periods [DONN79].
40
Final ly we a r r ive a t the m e c h a n i s m e m p l o y e d w h e n a t r u e
col l i s ion of d a t a occurs on the t runk . This is de t ec t ed when an
adapter fai ls to receive a response frame from a f r a m e t r a n s m i s s i o n
w i t h i n t i m e i n t e r v a l equal to the a d a p t e r ' s f ixed delay. I f this
s i t u a t i o n occurs , t he a d a p t e r s imply r e t r i e s t he t r a n s m i s s i o n .
R e s e r v a t i o n reques t f r a m e s are repeated 16 times on each trunk and
all other frames are retransmit ted 256 times.
A c o n c l u d i n g note to the HYPERchannel transmission protocol is
the messages can be transmitted in one of two ways: intra-adapter or
in te r -adap te r . A message t ransmission between devices linked to the
same adapter is called an intra-adapter transmission and t a k e s p l a c e
e n t i r e l y l o c a l l y w i t h no t runk t r a n s f e r . A t r a n s m i s s i o n be tween
devices on d i f f e r en t adapters is termed inter-adapter c o m m u n i c a t i o n .
A l t h o u g h i n t r a - a d a p t e r c o m m u n i c a t i o n s occur o f f - t r u n k , t h e s e
t rans fe rs also d i rec t ly a f f e c t the network throughput and trunk usage
f i g u r e s . S i n c e a d a p t e r s m a y b e e n g a g e d i c o n e a n d on ly o n e
c o m m u n i c a t i o n s e s s i o n a t a n y t i n e , i f a n a d a p t e r r e q u e s t s t h e
s e r v i c e s o f a n a d a p t e r e n g a g e d i n i n t r a - a d a p t e r b u s i n e s s , t h e
requested adapter will appear busy, as if it were engaged in a n o t h e r
t runk t r a n s f e r a c t i v i t y . I f , h o w e v e r , resources do no t c o n f l i c t ,
intra-adapter act ivi t ies may occur c o n c u r r e n t l y w i t h i n t e r - a d a p t e r
activi t ies.
Obviously, as the m e s s a g e g e n e r a t i o n andrecept ion needs of the network components change,or the network components themselves c h a n g e , thene twork th roughput and trunk usage f igures wil lchange. What is needed is a method for examiningt h e e f f e c t s o f c h a n g e s i n t h e s y s t e mconf igura t ion and this is the j u s t i f i c a t i o n forthis modell ing and s imula t ion e f f o r t .
SECTION 3
HOSC MODEL DESCRIPTION
3.1 The Real System
The HOSC is, as has been previously described, a distributed
computer network. This network is used to analyze and disseminate
information that is of interest to scientists and engineers at the
Marshall Space Flight Center as they monitor the progress of NASA
Space Shuttle missions. The primary components of the HOSC system
are the computers that perform the data manipulation tasks and the
HYPERchannel high speed local network that provides the dissemination
services. As the Space Shuttle mission priorities change from flight
to flight, so do the operating parameters of the Huntsville Operation
Support Center. In an effort to predict the effects of these changes
on the HOSC network, a model and a simulation of the HOSC were
developed.
3.2 Experimental Framework for the Case Model
The simulation task was accomplished by developing a general
model of a HYPERchannel network and by imposing parameters on this
model which describe the operations of IIOSC. These constraints form
an experimental frame through which the real system is observed and
exercised. The intent was to produce a relatively simple model that
would produce replicatively and predictively valid data corresponding
42
with the actual HOSC network. The term replicatively valid indicates
that the model data should match the data obtained from the real
system operating under conditions that match the parameters of the
simulation. Predictively valid implies that the results of the model
could be used to accurately predict the performance of the real
system in a hypothetical configuration. To achieve this model
validity, an attempt was to also make the model loosely structurally
valid [ZIEG76]. The modifier 'loosely' is applied to indicate that
although the model was not developed to correspond directly to the
hardware functions at the electrical and mechanical level of the real
system, the model was designed to correspond as nearly as possible to
the functional characteristics of the real system at the protocol
level. Relating this in some way to the proposed standard, seven
level, ISO-OSI model for distributed computer processing (Appendix
E), the model could be thought of as having been developed between
levels three and four, the Network and Transport layers, as opposed
to having been developed at the Physical layer, level 1.
The performance of the network model will be reported, using the
following measures:
o D: The delay that occurs between the time a framesequence becomes ready to transmit and the actualcompletion of the transfer.
o S: The throughput of the network reported as thetotal rate of data transmitted over the network.
o U: The utilization of the network normalized to one.This figure represents the fraction of the totaltrunk capacity used.
o G: The offered trunk load. This is the total rateof data presented to the network includingcontrol frames and retransmission frames.
43
o I: The input load. This is the rate of data that isg e n e r a t e d by an a d a p t e r a t t ached to the trunk.This represents only the user-generated da ta thatis t r a n s m i t t e d over the trunk and thus excludesthe control and retransmission f rames .
W h e r e poss ib le , these f igures wil l be reported for the network as a
whole, the individual trunks in the network, the individual adapters ,
and the individual devices.
3.3 Informal Descript ion of the Base Model
Components
TRUNK: C o a x i a l Cab le i n t e r c o n n e c t i o n H Y P E R c h a n n e lusers .
ADAPTER: Por t t h rough w h i c h a u s e r g a i n s a c c e s s to atrunk.
DEVICE:HYPERchannel user. The d e v i c e is the f u n d a m e n t a lent i ty of the network.
OFFNET-SOURCE: Non-network supplier of d a t a to a d e v i c e .Supp l i e s t h a t raw d a t a that will be reformedby the devices and routed over the ne twork .
FRAME-SEQUENCE: Basic quantity of in fo rmat ion that can bet ransported on the ne twork .
Descriptive VariablesTRUNK(i) (i=1.4)
SUCCESSFUL-TRANSFERS: R a n g e of n o n - n e g a t i v en u m b e r s i n d i c a t i n g n u m b e r o f b i t ssuccessful ly t r ans fe r red on the trunk.
OFFERED-TRUNK-LOAD: R a n g e o f n o n - n e g a t i v ec h a n n e l f o r t r a n s f e r . I n c l u d e s c o n t r o land retransmission sequences.
INPUT-LOAD: R a n g e o f n o n - n e g a t i v e n u m b e r s .Total number of user generated da ta b i t s o f f e r e dto the trunk for transfer.
TOTAL-DELAY: Propagation d e l a y of the t r unk infeet per microsecond ( f t / u s e c ) .
LENGTH-OF-TRANSMISSION-FRAME: Number of b i t s ina t runk t r a n s m i s s i o n s e q u e n c e . (Range from 64bytes to 4 ki lobytes.)
ADAPTER(i) ( i=l ,N)
44
SUCCESSFUL-TRANSFERS: Range of n o n - n e g a t i v enumbers ind ica t ing number of bits successfullytransferred on the trunk.
OFFERED-TRONK-LOAD: R a n g e o f non -nega t ivenumbers indicating number of b i t s s u c c e s s f u l l ytransferred on the trunk.
INPUT-LOAD: Range of n o n - n e g a t i v e n u m b e r s .To t a l number of user generated data bi ts offeredto the trunk for t ransfer .
DISTANCE-FROM-REFERENCE-ADAPTER: Number of f ee tof trunk cable of the a d a p t e r f r o m a t r u n kr e f e r e n c e adap t e r . Used for computing adapterdelay t imes.
PRIORITY-DELAY: Time (in usec ) indicating therelative transmission priority of the adapter.
PORT-STATUS: I n d i c a t i o n of whether an adapterport is supporting an active hose.
DEVICE(i) (i=1.4)TRANSMISSION-FREQUENCY: Probability dis t r ibut ionindicating the relat ive frequency that the devicewill request the trunk (range from 0 to 1.0).
INTERNAL-BUFFER-SIZE: N o n - n e g a t i v e n u m b e rindicating the amount of data b u f f e r e d by a hos tbefore requesting the trunk.
RETRY-COUNT: In t ege r in the r a n g e of 1 to 64ind ica t ing the number of t i m e s an adapter wi l lrepeat a reservation request ,
I/0-BUS-TRANSFER-RATE: Real number in units ofkilobytes per second giving the speed w i t h w h i c hda ta may be t r a n s f e r r e d f rom the host processorto the adapter.
SUCCESSFUL-TRANSFERS: R a n g e of n o n - n e g a t i v enumbers indicating number of b i t s s u c c e s s f u l l yt ransferred on the trunk.
OFFERED-TRUNK-LOAD : R a n g e of n o n - n e g a t i v enumbers . To ta l number of b i t s o f f e r e d to thechannel f o r t r a n s f e r . I n c l u d e s c o n t r o l a n dretransmission sequences.
INPUT-LOAD: Range of n o n - n e g a t i v e n u m b e r s .Total number of user-generated data bi ts o f f e rdto the trunk for transfer.
45
SOURCE(i) ( i=l.N)DATA-STREAM-RATE: R e a l n u m b e r i n u n i t s o fk i l o b y t e s per second i nd i ca t i ng the continuousrate at which data arrives at a d e v i c e f r o m theof fne t source.
TRANSMITTING-DEVICE: The i n i t i a t i n g d e v i c e e n g a g e d in atransmission session.
RECEIVING-DEVICE: A non- in i t i a t ing d e v i c e e n g a g e d in atransmission session.
PARAMETERSBANDWIDTH: Bandwidth of network trunk (50 Jib itsper second).
TIME: Non-negative real number representing taeinterval of time during which the mode l w i l l bevalid.
Component In te rac t ion •FRAME-SEQUENCES are t ransmit ted over a TRUNK from DEVICE toDEVICE through the network ADAPTERS. The number of FRAME-SEQUENCEs in a transmission session is determined from theINTERNAL-BUFFER- SIZE of the DEVICE. The f r e q u e n c y oft r a n s m i s s i o n s e s s i o n s is de t e rmined from the SOURCE DATA-STREAM-RATE and the DEVICE TRANSMISSION FREQUENCY.
SUCCESSFUL-TRANSFERS o c c u r if the FRAJIE-SEQUENCE ist ransfer red from the TRANSMITTING-DEVICE to the RECEIVING-DEVICE w i t h o u t co l l i s i on . If the TRUNK is engaged in at ransmiss ion when a DEVICE becomes r e a d y to t r a n s m i t , theADAPTER w i l l be d e l a y e d u n t i l the PRIORITY-DELAY expiresfor the ADAPTER and the TRUNK is f r ee .
SECTION 4
HOSC SYSTEM SIMULATION DESIGN
4.1 Scope
This section describes a computer simulation developed from the
system model for the Huntsville Operation Support C e n t e r . The goal
of the s imulat ion is to provide a sof tware tool which may be used to
study the operating c h a r a c t e r i s t i c s of the HOSC. The tool should
h a v e t h e c a p a b i l i t y o f g e n e r a t i n g d a t a t h a t r e p l i c a t e s t h e
performance of the exist ing HOSC system, as well as the capabil i ty of
p r e d i c t i n g the p e r f o r m a n c e o f t he HOSC s y s t e m to h y p o t h e t i c a l
configurations.
4.1.1 Major Software Functions
The primary functions of the sof tware should be to:
1. Provide a repl icat ively and pred ic t ive ly valid s imulat ion of the
real system model described in Section 3,
2. P rov ide an easy method for modifying the s imulat ion environment
so that it may be used to s imulate e n v i r o n m e n t s d i f f e r e n t f r o m the
existing system,
3 . Provide m e a n i n g f u l o u t p u t t h a t m a y b e c o n v e r t e d t o t h e
performance indicators described in Section 3 and 5,
4 . Provide a user f r i e n d l y i n t e r f a c e a t a l l l e v e l s of u s e r
interaction.
47
The software should be designed:
1. To be as portable as possible so that the simulation may be
transferred easily from the development computer to other computer
environments.
2. To allow enhancements of the fundamental algorithm as time and
further study permit.
4.1.2 System Files
The HOSC s o f t w a r e simulation requires the maintenance of three
f i les for proper execution; the program f i l e contains the ob jec t code
f o r t h e s i m u l a t i o n p r o g r a m , a n d t w o a u x i l i a r y d a t a f i l e s s t o r e
information that is used by, or created by, the s i m u l a t i o n p r o g r a m .
The aux i l i a ry da ta f i l e , DESCRIP-FILE, is used for long term s torage
o f d a t a used t o d e s c r i b e t he s y s t e m to be s i m u l a t e d . A t t he
b e g i n n i n g o f e a c h s i m u l a t i o n , t he use r i s g i v e n the o p t i o n o f
simulat ing the system described in DESCRIP-FILE or of c r e a t i n g a new
s y s t e m d e s c r i p t i o n whi le on-line wi th the s i m u l a t i o n program. The
auxil iary f i l e , AUX-OUT, i s m a i n t a i n e d t o t a l l y by the s i m u l a t i o n
s o f t w a r e and p rov ides long term storage for the d e t a i l e d summary of
the simulation. This f i l e is pr int-fornat ted and may be d i r e c t e d to
a system printer at the user ' s request .
M o d i f i c a t i o n o f t he s y s t e m d e s c r i p t i o n f i l e , DESCRIP-FILE,
a l l o w s the user to change to the s i m u l a t i o n s c e n a r i o and can be
changed during the characterizat ion stage of each run . D u r i n g th i s
s t a g e of the run, the user i s g iven the o p p o r t u n i t y to p r o v i d e
48
additional input for the simulation run through interactive queries.
These queries establish such run dependent parameters as the seed for
the random number generator, descriptive titles for the detailed
output summaries, and the interval of simulated time for which the
simulation run will execute.
Both auxiliary files are maintained as standard ASCII text
files. The program file may be maintained as either source or object
code since code modification should be unnecessary except during
algorithm enhancements.
4.1.3 Major Design Constraints
The primary design constraint imposed on this simulation effort
was the the simulation be transportable between computer systems of
unspecified size and manufacturer. To accommodate this constraint,
the use of a large, general purpose simulation language, such as SLAM
or GPSS, was considered impractical. Therefore, the language Pascal
was chosen because of its widespread availability and its richness of
data structures.
4.2 Design Specifications
4.2.1 Derived Software Structure
The fundamental system model (FSM) for the simulation program,
shown in Figure 4.1, illustrates the interfaces between the
simulation program, the user, and the two auxiliary files.
The FSM may be decomposed into three separate a c t i v i t i e s :
characterization, simulation, and summary. This composition leads to
USER
DetailedDescription of
Network
Request for Simulationand Specificationof Parameters
Descrip-File
49
USER
Brief Summary ofSimulation Results
Aux Out
Figure 4.1 Full Scale Model
USER
Descrip-File
Aux-Out
Figure 4.2 First Level Refinement
50
the possible refinement of the FSM into the data flow diagram of
Figure 4.2.
The Characterization module has a dual task; initialization of
the run parameters, and specification of the experimental framework
of the simulation. The Summary module may also be divided into two
distinct activities: a brief summarization of the network statistics
and a more detailed reporting of the simulation activities. This
breakdown suggests the refinement shown in Figure 4.3.
The Simulation activity can be described by the following
algorithm which increments the simulation clock each time a
significant event occurs in the simulation.
Set Clocks at Beginning Tine
Find Transmitter
Determine a Receiver
Attempt to Transmit
If No Collision then
Clocks may be Updated toreflect successful transmission
Find next Transmitter
else
Clocks must be updated toreflect collision
Find next Transmitter
Repeat until simulation complete
With this algorithm, the data flow diagram shown in Figure 4.4may be used to describe the simulation activities.
51
cx
c0)
0)c
•HU-l0)
OS
TJ
oaoc/:
<l0)
3ft.
Io
cc
52
c01
<uc•Hu-^OJas
01>0)
<rOii-ec•HPI-
53
4.2.2 Interfaces within the structure
MODULE INPUT OUTPUT
VARIABLES VARIABLES
Initialize none
CharacterizeNetwork
None
PrintNetDescription Adapter arrayPr
FindNext
PicA
ACollision
UpdateClocks
PrintNetworkStats
Adapter array
TisitterPrAdapter array
AdapterTmitter
Tmitter
TxTallyColiisionTallyWaitTallyTo talAttenptsTotalLostTimeActive
HeadingMaxTimeMaxTxU, BMC seedPrintAllInputMode
Adapter arrayPr, Transmis. Prob,NumofTrunksTrunkLength array
none
ACollision
TiueBit sixTxTallyNextTxTsCtTxTitieRxCtRxTiineTineActive
54
MODULE INPUT OUTPUTVARIABLES VARIABLES
ActivitySumnary Adapter noneTiTallyCollisionTallyWaitTallyTotalAttemptsTotalLostTimeActive
55
4.3 Module Descriptions
4.3.1 ModSc
4.3.1.1 Processing Narrat ive
Mod3c is the top level module of the HOSC s i m u l a t i o n p r o g r a m .
I t d e f i n e s the s i m u l a t i o n env i ronment and p r o v i d e s the f l o w o f
con t ro l n e c e s s a r y f o r t h e s i m u l a t i o n . M o d S c c a l l s p r o c e d u r e s
I n i t i a l i z e a n d C h a r a c t e r i z e N e t w o r k t o c r e a t e a n d d e f i n e t h e
s i m u l a t i o n e n v i r o n m e n t . U s i n g t h i s c h a r a c t e r i z a t i o n o f t h e
e n v i r o n m e n t . ModSc c o n t r o l s the a c t u a l s y s t e m s imula t ion as was
i l lustrated in the data flow diagram, Figure 4.4.
At the c o m p l e t i o n of the s imula t ion ac t iv i ty , ModSc i n i t i a t e s
procedures PrintNetworkStats and ActivitySummary. P r i n t N e t w o r k S t a t s
provides a brief summary of the s imulat ion and pr in t s this summary in
the auxil iary output f i l e and on the s y s t e m d e f a u l t o u t p u t d e v i c e .
A c t i v i t y S u m m a r y compiles a deta i led recount of the s imula t ion at the
individual device level and r e c o r d s th is summary in the a u x i l i a r y
output f i l e .
56
4.3.1.2 Interface Description
Procedure call: program ModSc (Input.Output.DescripFile.AuxOut)
Parameters: Input: system defined default input fileOutput: system defined default output fileDescripFile: Text (* File of Char *)AuxOut: (Text (* File of Char *)
Called by: none — Top Level
Calls: InitializeCharacterizeNetworkPr in tNetDe s c r ip t ionFindNext(Tmittr)PickA(Rcvr)ACollisionUpdateClocks(Condition!])PrintNetworkStatsActivitySummary
Global variables referenced:
Current TimeMaxTime
JiaxTx
TxTally
CollisionTallyTotalAttempts
TmittrRcvr
Condition
(* Simulation time *)(* Simulation parameter.
When CurrentTime>MaxTime,simulation terminates *)
(* Simulation parameter.When TxTally > MaxTx.simula/tion terminates *)
(* Sum of successful *)transmit ted)
(* Sun of collisions *)(* Sum of transmissions
and collisions *)(* Active Transmitter *)(* Active Receiver, object
of transfers fromTmittr *)
(* Indicates whether or notcollision occurred intransmission *)
57
4.3.1.3 Design DescriptionData Structure
typeAdapterRecord = record
Device: array[1..4] of Device Records;PriorityDelay: real;EndDelay: real;
DistFromAl: real;ATxCT: integer; (* Adapter Transmission Tally *)AWaitCt:integer; (* Adapter tally of Waits *)ACollCt:integer: (* Adapter tally of Collision *)ATxTiine: real; (* Adapter time spend in Transm.*)AWaitTime:real; (*Adapter time spent waiting *)ACollTiue:real; (*Adapter time spend in Colls. *)
end;
DeviceRecord = recordOpen: boolean (* Flag signaling port status *)if NOT Open thenID: string of char; (* Device name *)RetryMax: integer; (* Retry Count *)TferRate: real; (* I/O port transfer rate *)
NumofSources: integer; (*Num of cffnet sources*)Source: array[l. .MaxNumSources] of
SourceRecords;LoadTine: real (* Inverse of Tfer Rate *)
end;
SourceRecord = recordGenRate: real; (* Data generation rate Bytes/s *)BufferSize: integer; (* Size of Device buffer,
Amount of data ace urn. before tx *)ID : string of char;OnTrunk: array[1..4] of boolean; (*List of trunks
to try *)NextTx: real; (*Time of next transin. *)LastTx: real; (*Time of last transm. *)Txlntrvl: real; (*Time between transit. *)TxCt: integer: ("Tally of transm. *)WaitCt: integer; (*Tally of Waits *)CollCt: integer; (*Tally of Collisions *)RxCt: integer; (*Tally of times as receiver *)TxTime: real; (*Time in transm. *)WaitTime: real; (*Time in Waiting *)CollTime: real; (*Time in Collisions*)RxTime: real; (*Time as Receiver *)
end;
58
Variables
Adapter: array[1..MaxNumAdapters] of AdapterRecord;Pr: array[l..NumOfDevices,!..NumOfDevices] of real;
(* Pr is probability of Transmission matrixstored as cummulative probabilities *)
Tmittr, Rcvr: SourceDescriptions;
Algorithm
begin
Initialize variables
Characterize network
Print description in auxiliary output file
Set clock at beginning time
Find Transmitter
Determine Receiver
Attempt to transmitif No Collision then
Clocks may be updated toreflect successful transmission
Find next transmitter •else
Clocks must be updated toreflect collision
Find next transmitter
Repeat until simulation complete
Print Summary of Network Statistics
Print Activity Summary to auxiliary output fileend.
59
4.3.1.4 Modules Used
InitializeCharacterizeNetworkPrintNetDescriptionFindNextPickAACollisionUpdateClocksPrintNetworkStatsAc t iv i tySumma ry
Descriptions of these modules will be contained in the fol lowing
sections.
4.3.2 Initialize
4.3.2.1 Processing Narrative
Procedure Initialize is invoked by the top level module, Mod3c,
a t t h e o u t s e t o f e a c h s i m u l a t i o n r u n w i t h t h e p u r p o s e s o f
i n i t i a l i z i n g the global program variables, and of establishing other
p a r a m e t e r s fo r t he run. V a r i a b l e i n i t i a l i z a t i o n s a re p e r f o r m e d
w i t h o u t user-intervention with the purpose of giving each var iab le a
de f ined and c o n s t r a i n e d va lue a t the b e g i n n i n g of the r u n . Tlie
g l o b a l p a r a m e t e r s t h a t c a n n o t b e a d e q u a t e l y d e f i n e d t h r o u g h
initialization are assigned values in response to interact ive queries
posed to the use r . (A l i s t of the interact ive queries is shown in
Table 4.1. Also included in the table are some of the c o n s t r a i n t s
imposed on the responses to these quest ions.)
60
TABLE 4.1 INTERACTIVE QUERIES POSED BY INITIALIZE
Query: Identification Heading for run:Response Valid responses are alphanumeric strings
with length less than SO characters. Thisheading appears as a heading for the detailedoutput summaries.
Example: Simulation Configuration la — 4/24/84Variable: Heading
Query: Maximum tun tine in sect?Response: Valid responses are real numbers specifying
the interval of simulated time for which thesimulation will run.
Example: 60.0Variable: MaxTime
Query: Maximum •uooettful transmissions?Response: Valid responses are integer numbers setting
the simulation interval in terms of the numberof successful transmissions that occur of thenetwork. Simulation is terminated whentransmission count exceeds this value or whenmaximum run tine has been exceeded.
Example: 500Variable: MaxTx
Query: Seed for random number generator?Response: Integer number. For best results, the number
specified should be an odd, positive number withat least 5 digits, but not divisible by 5 or 7.
Example: 316227Variable: U
Query: PRINT-ALL condition on?Response: Y or N. if Y, then a record of each network,
activity will be printed on the system's defaultoutput device. Due to the magnitude of outputgenerated, this query should be answered Y forshort simulation runs only.
Example: NVariable: PrintAll and Resp (See Comments)
Query: Input mode: INTERACTIVE or FILEResponse: I or F. If File is specified, network
description will come from DecripFile,otherwise, the interactive network specificationwill be initiated.
Example: FVariable: Input Mode and Resp (See Comments)
61
4.3.2.2 Interface Description
Procedure call: Initialize
Parameters:
Called by:
Calls:
none
ModSc
none
Global variables referenced:
WaitTally
CollisionTally
TotalAttempts
Pet
TiceActive
BitsTx
Heading
MaxTime
MaxTs
U
PrintAll
InputMode
4.3.2.3 Design Description
(*Tally of devices putinto waiting states duringtrans attempts*)(*Tally of collisionsoccurring on network *)(*Sum of waits, colls.,and transmissions *)("Tally of pages inoutput file *)transmitted on network*)(*Total time duringwhich trunk in active*)(*Total number of bitstransmitted on network*)(•Identification headingfor the run *)("Interval of simulatedtime for which sim. willcontinue *)(*Number of transm. forwhich siniul. will run *)(*Seed for random numbergenerator *)(*Flag for detailedtrace of simulation *)(^Variable indicatingsource of network desor.File or Interactive *)
This procedure is simple straight line code containing the
statements necessary to make the desired i n i t i a l i z a t i o n s and
assignments.
4.3.2.4 Modules Used
None
62
4.3.3 CharacterizeNetwork
4.3.3.1 Processing Narrative
Procedure Charac ter ized twork is invoked by the top l e v e l
m o d u l e , M o d S c , w i t h the purpose of e s t a b l i s h i n g the experimental
framework of the simulation. In this p r o c e d u r e , the p r imary d a t a
s t r u c t u r e s representing the abstraction of HOSC network are assigned
values. These values are defined either interactively by the user at
the time of simulation, or they may be called from the auxiliary data
f i le . DescripFile. The choice of origin of the s y s t e m d e s c r i p t i o n
is controlled by the variable InputMode, which is assigned a value in
procedure Initialize. If InputMode has the va lue In te rac t ive . the
s y s t e m descr ip t ion is controlled by the variable InputMode, which is
assigned a value in procedure Ini t ia l ize . If InputMode has the value
In t e rac t ive, the s y s t e m d e s c r i p t i o n is per formed interact ively and
stored in DescripFile. An InputMode va lue of F i le lnput s p e c i f i e s
t h a t the network description will be read from DescripFile, implying
a previous run in which DescripFile was created interact ively.
Network characterization is accomplished in two s tages : ne twork
desc r ip t i on and p r o b a b i l i t y - o f - t r a n s m i s s i o n e s t a b l i s h m e n t . A
d e t a i l e d l i s t i n g o f t h e i n t e r a c t i v e q u e r i e s posed du r ing t h e
characterization process is presented in Table 4.2.
F o l l o w i n g t h e c h a r a c t e r i z a t i o n o f t h e n e t w o r k , t h e
probability-of-transmission matr ix is enumerated. This e n u m e r a t i o n
i s p r o m p t e d by the q u e r y , ' C O M M U L A T I V E PROBABILITY MATRIX' ,
whereupon, the probabilities will then be c o n v e r t e d to c u m n u l a t i v e
values and stored in the matrix, PR.
63
TABLE 4.2. PROMPTS FOR NETWORK CHARACTERIZATION
QUERY: Total number of trunks in network:(Maximum number of n)
Response: Integer number in range 1 to MaxNumTrunks.Variable: NuiaofTrunks
Query: Length of trunks in feed (used for END DELAY)Trunk 1:Trunk 2:
Trunk n:Response: Positive real numbersVariable: TrunklengthU]
Query: Number of Adapters in Network:Response: Integer number in range 1 to MaxNumAdaptersVariable: NumofAdapters
Query: Adapter #i:Distance in ft from common trunk Adptr 1:
Response: Positive real number. Establishes a reference adapteron each trunk.
Variable: Adapter[i].DistFronAl
Query: Adapter Priority delay:Response: Positive real number representing seconds of delayVariable: Adapterli].PriorityDelay
Query: Adapter[i].Devioe[j]desoriptionDevice status (OPEN or CLOSED):
Response: 0 or C. 0 if adapter[i], device porttjj lias anactive, network device connected to it. C otherwise.
Variable: Adapter[i].Device[j].Open
Query: Device ID «-24 char)Response: Alphanumeric stringVariable: Adapterli].Device[j].Id
Query: Device to Adapter I/O rate (Bytes/s):Response: Positive real number representing the speed of the
I/O port of the device.Variable: Adapter [i].Device[j].TferRate
Query: Number of data sources in DeviceResponse: Integer number representing the number of off-net
sources supplying device with data.Variable: Adapter[i].Device[j].Source[k].Id
64
Query: Buffer size in Byte*:Response: Integer number. Size of buffer filled in device
before network transmission is requested.Variable: Adapter[i].Device[j].Source[k].BufferSize
Query: Trunks *oo««*ed by Source[k](Bator I if connected, N if not)Trunk 1:Trunk 2:
Trunk 1:Response: Y or N. Indicates the trunks over which data from each
source will be transmitted.Variable: AdapterQ] .Oevice[j] .Source[k] .OnXrunk[l]
65
4.3.3.2 Interface Description
Procedure call: Characterized twork
Parameters: none
Called by: &od3c
Calls: RewriteDescripFileInitializeProbMatrix
Global variables references:
Adapter ('Records containing descriptionof each adapter in network *)
PR ('Probability of TransmissionMatrix *)
NumofAdapters ('Network adapter count *)NumofTrunks ('Network trunk count *)Dnum ('Network device count *)
Files accessed:
Input ('Default Input File*)Output ('Default Output File*)DescripFile ('Auxiliary Output File*)
4.3.3.3 Design Description
This procedure is designed using simple code containing the
statements necessary to make the desired queries and variable
initializations.
4.3.3.4 Modules Used
Rev/riteDescripFile — Subordinate procedure employed torewrite the auxiliary descriptionfile, DescripFile. Requires accessto the global variables. Adapter,PR, NumofAdapters, NumofTrunks, andDnum.
InitializeProbMatrix- Subordinate procedure employed toestablish the values of theprobability of-transmission matrix.PR. Requires access to globalvariables Adapter. PR, NumofAdapters,and Dnum.
66
4.3.4 PrinNetDescript ion
4.3.4.1 Processing Narrative
Procedure PrintNetDescription is invoked by the top levels
module, ModSc, with the purpose of printing a copy of the network
description in the auxiliary output file, AuxOut.
4.3.4.2 Interface DescriptionProcedure call: PrintNetDescriptionParameters: noneCalled by: Mod3cGlobal variables referenced:
Adapter (*Array of records containing thedescription of each adapter inthe network.*)
PR ('Probability of transmission matrix*)NumofAdapters (* Number of adapters in network *)NumofTrunks (* Number of trunks in network *)Dnum (*Number of devices in network*)Let (*Number of lines in AuxOut *)Pet (*Nuinber of pages in AuxOut *)
Files accessed:AuxOut ('Auxiliary output files*)
4.3.4.3 Design Description
This procedure is designed using simple code containing the
statements to copy the Adapter records into the auxiliary output
file, AuxOut. Adapter records are copied with an appropriated
heading of numbered pages in the file.
4.3.4.4 Modules Used
NewPage Procedure used to copy blank lines into thethe auxiliary output file at places appropriateto give page breaks. Procedure NewPage requiresaccess to the file, AuxOut and to the globalvariables Let, Pet, and Heading.
4.3.5 FindNext
4.3.5.1 Processing Narrative
67
Procedure F indNext is invoked by the top level module Mod3c,
with the purpose of choos ing an a p p r o p r i a t e d t r a n s m i t t e r in the
s i n u l a t e 4 d n e t w o r k . The s imula t ion process is event-driven w i t h a
device 's request for ne twork se rv ice ac t ing as the p r i m a r y e v e n t .
The simulation clock is advanced each time a significant event occurs
in the simulation. Although the clock may be advanced in cases of
s i m u l a t e d co l l i s ions and w a i t c o n d i t i o n s , the pr imary simulat ion
event is the c o m p l e t i o n of a s u c c e s s f u l t r a n s m i s s i o n . P r o c e d u r e
FindNezt examines the characteristics of each device, and chooses a
ready t r a n s m i t t e r ba sed on the amount of da t a tha t a d e v i c e i ray
r e c e i v e f r o m an o f f - n e t w o r k source and the b u f f e r s ize of tha t
source. This d e t e r m i n a t i o n is ba sed on the cu r ren t va lue of the
system clock and the time of the device ' s next transmission reques t .
4.3.5.2 Interface Description
Procedure call: FindNext(Tmit tr)
Parameters : Tmittr (*Variable of typeSourceDescription represent ingthe simulated ne twork ' s currentlyactive transmit ter*)
Called by: Mod3c
Calls: none
Global variables referenced:
Adapter , (*Array of records containingdescription of each adapter innetwork *)
NumofAdapters (*Number of adapters in thenetwork *)
68
4.3.5.3 Design Description
Algorithm
begin (* procedure FindNext *)
for I := 1 to Number of Adapters
for J: = 1 to 4 (* scan each of the 4 possible deviceson the adapter *)
for K: = 1 to Number of Sources
choose adapter with earliest timeof next transmission
end
end
end
return the attributes of the next transmitterthrough the procedure parameter
end (* procedure *)
4.3.5.4 Modules used
none
4.3.6 Picka
4.3.6.1 Processing narrative
P r o c e d u r e P i c k a examines a l l p o s s i b l e r e c i p i e n t s o f d a t a
genera ted by t ransmit ter that is currently ac t ive , and des igna tes one
of t hese dev ices to be the act ive receiver . This choice is made by
examining the probabili ty-of-transmission m a t r i x . PR, and a r a n d o m
number g e n e r a t e d by the pseudo-random number generat ing procedure.
Uniform. The values stored in PR represent cummulative probabi l i t ies
of t ransmiss ion from one device to another, so, the determinat ion of
the receiver is made by searching the row of the mat r ix r e p r e s e n t i n g
69
the possible receiving devices of the transmitter until the table
entry is larger than the random number.
4.3.6.2 Interface Description
Procedure call: Picka(Rcvr, Tmitter)Parameters: Rcvr ('Variable of type SourceDescrip-
tion assigned the attributes of thecurrently active receiving device.*)
Tmittr (*Variable of type SourceDescrip-tion assigned the attributes of thecurrently active transmitting device *)
Called by: Mode3c
Calls: Uniform ("Uniform random num generator*)
Global variables referenced:
Adapter (*Array of records containing thedescriptions of each adapter inthe network *)
NunofAdapters (*Nunber of adapters inthe network*)
U (*Real variable assigned the valueof the random number chosen byUniform *)
4.3.6.3 Design Description
Algorithm
begin (* procedure Picka *)
choose Uniform random number
determine receiver based on random number andprobability matrix
assign attributes of receiver to parameterReceiver
end (* procedure *)
4.3.6.4 Modules used
Uniform Uniform is a routine that calculates a uniformpseudo-random number in the range from 0 to 1. Theglobal variable, U, is accessed in the procedure andused as the seed for the random number calculation.After the first call to the routine, the seed becomes
70
the last random number calculated. The initialvalue of the seed should be an odd, positive numberof at least 5 digits, but not divisible by 5 or 7.
4.3.7 ACollision
4.3.7.1 Processing narrative
The purpose of the boolean function, ACollision, is to determine
whether or not the currently transmitting device has free and
unrestricted use of the simulated transmission medium. If another
device attempts to transmit within a length of time equal to, or less
than, the EndDelay of the trunk, a collision is likely to occur. The
purpose of ACollision is to compare the transmission time of the
currently active device with the NextTx time of each of the other
devices on the network. If another device has a NextTx time
scheduled to fall within EndDelay number of seconds from the current
time, then a collision will be declared and the value of ACollision
will be set to true. The function compares the transmission times to
a constant called CollEps, which should be assigned a value equal to
the EndDelay of the trunk./
4.3.7.2 Interface description
Function call: ACollision
Parameters: none
Called by: Mod3c
Calls: none
Global variables referenced:
NumofAdapters ('Network adapter count *)Adapter (*Array of adapter records
in which are stored theattributes of each recordin the network *)
71
CollisionTally ('Tally of collisions thatoccur on the network *)
PrintAll ('Program flag that, when setto True, specifies theprinting of each collisionfor a detailed trace of thesimulation*)
4.3.7.3 Design description
Algorithm
begin (* function ACollision *)
for I := 1 to NumofAdapters do
for J := 1 to 4 do (* Examine each port *)
for K: = 1 to NumofSources do
if abs(Device.NextTx - Transmitter.NextTx)
<= CollEps then
set function value to True
end
end
end
end (* function ACollision *)
4.3.7.4 Modules used
none
4.3.8 UpdateClocks
4.3.8.1 Processing narrat ive
Procedure Upda tec locks is invoked by the top- leve l m o d u l e ,
M o d 3 c . U p d a t e C l o c k s i s the mos t c r u c i a l l ink in the s i m u l a t i o n
p r o g r a m s ince th i s i s the m o d u l e in w h i c h the b u l k of the
72
HTPERchanne1 p r o t o c o l is m o d e l l e d . As po in ted oat in p rev ious
sections, the simulation program is event-driven, indicating that the
s i m u l a t i o n c lock is advanced each time a significant event occurs.
This me thod of d r iv ing the s i m u l a t i o n i s o f t e n r e f e r r e d to as
activity scanning, or event scanning.
The e v e n t t h a t i n i t i a t e s the p r o c e d u r e i s a s i m u l a t e d
t r ansmiss ion request. Each device (or ent i ty) is assumed to possess
certain transmission a t t r ibutes determined from obse rva t ions of the
r e a l s y s t e m . For example , i f a device r e c e i v e s da ta for t runk
t r a n s f e r f rom an o f f - n e t w o r k s o u r c e a t an a v e r a g e r a t e of 15
k i l o b y t e s per second, and stores that data in a 2096-byte bu f f e r , -it
will request a trunk transmission about once every 0.14 seconds. (At
t h e p r e s e n t l e v e l o f d e t a i l , t h e m o d e l d o e s n o t accoun t f o r
reservation requests made by adapters wish ing to t r a n s m i t on the
t runk, t h e r e f o r e , the simulation assumes that a trunk transfer is a
transfer o f a c tua l d a t a . ) This t r a n s m i s s i o n r e q u e s t b e c o m e s the
event that drives the simulat ion clock.
The clock will be updated in one of several ways , depend ing on
the circumstances surrounding the requested transmission. If another
dev i ce w i s h e s to t r a n s m i t a t the same t i m e , o r w i t h i n E n d D e l a y
seconds of the cur ren t t ime , then a c o l l i s i o n w i l l occur and the
clock w i l l b e u p d a t e d t o s i m u l a t e t h e r e s p o n s e o f t h e r e a l
HYPERchannel system to a collision. The real system response to this
si tuat ion is to repeat the transmission at tempt up to a to ta l of 256
r e t i r e s . I n t he r e a l s y s t e m , th is i s an a d e q u a t e m e t h o d fo r
73
separating the conflicting adapters since a certain amount of time is
required for each adapter to respond to the detected collision. This
response time will probably be different for each adapter based on
the random differences in hardware. Even in the worst case, where
two adapters attempt to reserve each other at exactly the same time,
the adapters would probably be able to separate themselves in the 256
retries. The simulation program must respond to a collision in a
slightly different, and more artificial, way. In the event of a
collision, the NextTx time of each of the'conf1icting devices is
advance by a random percentage of the trunk EndDelay. With this
algorithm, the conflicting adapters may be separated.
If a device becomes ready to transmit during the time that
another adapter has control of the transmission medium, the ready
adapter's NextTx time is advanced by an amount equal to the adapter's
DelayTinie. This is, therefore, the method used by DpdateClocks to
simulate the wait state of the real system. The third method for
updating the clock is the simulation of a successful trunk
transmission. In this case, the simulated system clock is advanced
by the amount of time required for the block of d a t a to be
transferred over the trunk. Any device that becomes ready to
transmit within this interval will, of course, be forced into a wait
state.
As can be inferred from the preceding discussion, parts of the
model are rather naive abstractions of the real system. This is
especially evident from the fact that the simulation does not attempt
74
an abstraction of the reservation rejection problems described in the
detailed description of the HYPERchannel protocol of Section 2.2.
4.3.8.2 Interface description
Procedure call:
Parameters:
Called by:
Calls:
UpdateClocks(Condit ion[])
Condition
Mod3c
(*Var. of Normal, Collision,or Other. Normal indicatesno collision *)
Transfer!ime(Tmittr. Rcvr)( 'Calculatesto transmit
PrintWaitStats
PrintLostStats
time requiredentire data
block between adapters *)('Utility routine thatprints detailed recordof wait state *)('Utility routine thatprints detailed recordof lost messages*)
Global variables referenced:Tmittr
Rcvr
CurrentTimeBitslx
TxTallyPrintAll
NumofAdaptersWaitTallyTimeActive
('Currently activetransmitting device *)('Currently active receivingdevice *)('Simulation time *)('Total bits transm. ontrunk*)('Transmission Count *)('Flag for detailed trace ofsimulation *)
('Adapter count *)('Wait state count *)('Tine trunk involved intransmissions *)
75
4.3.8.3 Design description
Variables
DelayTime: real (*Amount of time device to be delayedin event of collision*)
Time: real (^Auxiliary storage for time to transdatabetween devices*)
Algorithm
begincase COND of
Normal (* No collision *): begin
Calculate Time by invoking TransferTime
Increment BitsTx (*Total bits transferred*)
Advance clock
Tally a successful transmission
Check to see if another adapter becomesready during transfer intervalif true then (* initiate wait state *)begin
Calculate Total Adapter Delay
Advance adapter's NextTx by Total Delay/
Tally wait conditionsend
Advance transmitter's NextTxend (* Successful Transmission case *)
Collision: beginTally collision statistics
Compute time to delay for collision
Increment Retry tally
Repeat until RetryTally > 256
If RetryTally > 256 then tally lost messageend (* Collision case *)
Other; endend (* case *)
end (* Procedure UpdateClocks *)
76
4.3.8.4 Modules used
TransferTime — Calculates the time requires to t ransmit amessage from a sender to a receiver. Passesparameters for t ransmit ter and receiver .
Uniform — Calculates a uniform pseudorandom number inthe range 0 to 1. Parameter is seed for theRNG.
4.3.9 PrintNetvorkStats
4.3.9.1 Processing narrative
P r o c e d u r e P r i n t N e t w o r k S t a t s p r i n t s a b r i e f summary of the
s i m u l a t i o n a c t i v i t y a t t h e e n d o f t h e s i m u l a t i o n r u n . T h e
a b b r e v i a t e d summary i s d i r e c t e d to t he s y s t e m ' s d e f a u l t o u t p u t
device, and to the auxiliary data f i l e , AuxOut . The f o r m a t of the
brief summary is as fol lows:
***END OF RUN***
Current time: xxxxxxx.xxxx sees
Successful t ransmissions: nnnnn
Collisions : iinnnn
W a i t s : nnnnn
Total Attempts : nnnnn
Total messages lost : nnnnn
Total time active : xxxxxx.xxxx
% to ta l time active : xxx.xx
Total Bytes t ransmit ted : nnnnnnnnnnnn
Seed for RNG : nnnnnnn
77
4.3.9.2 Interface description
Procedure call: PrintNetworkStats
Parameters: none
Called by: Mod3c
Calls: none
Global variables referenced
CurrentTime (* Simulation clock *)TxTally (* Successful transm. *)CollisionTallyWaitTallyTotalAttempts (*Sum of transmissions,
collisions, and waits *)TotalLost (* Tally of lost transm. *)TimeActive (* Time trunk spend servicing
devices *)PerCentActive (* TimeActive/CurrentTine *)BitsTx (* Total bits transm. *)lu (* Storage for random number
seed *)Let (*Line count in AuxOut *)
Files accessed:
Output (* Default output file *)AuxOut (* Auxiliary output file *)
4.3.9.3 Design description
This procedure .is designed using simple code containing the
statements necessary to write the desired variable values in the
auxiliary output file in the default output file.
78
4.3.9.4 Modules used
none
4.3.10 ActivitySummary
4.3.10.1 Processing narrative
Procedure ActivitySummary produces a detailed summary of the
simulation activities and prints this summary in the auxiliary output
file, AuxOut. The summary is compile at the device label, with a
tally of the activities of each device included in the summary of
tables. In addition to the simple tallies of transmissions,
collisions and waits, a separate tale is compiled containing the
percentage times that the devices spend engaged in each activity.
4.3.10.2 Interface description
Procedure call: ActivitySummary
Parameters: none
Called by: Mod3c
Calls: NewPage
Global variables referenced: .
NumofAdapters ("Number of adapters in thenetwork *)
Adapter ("Array of records in whichattributes of each adapterof each adapter stored *)
TimeActive ("Total time trunk spend inservicing adapter requests
Let ("Line count of AuxOut *)
Files Accessed:
Output (* Default Output file *)AuxOut (* Auxiliary Output file *)
79
4.3.10.3 Design description
Procedure ActivitySummary is primarily a format t ing routine and,
as such, l i t t le process ing is done. P e r c e n t a g e - a c t i v e t imes are
c a l c u l a t e d in two ways: percentage of the time the device is active
on the trunk, and percentage of the total time the trunk is active.
4.3.10.4 Modules used
NewPage — Procedure NewPage places page breaks at appropriate
places in the auxiliary output file. AuxOut.
4.5 Validation Criteria
The v a l i d a t i o n of th i s model w i l l depend p r i m a r i l y on i t s
s t r u c t u r a l v a l i d i t y . A l t h o u g h , t h e m o d e l l e d s y s t e m e x i s t s
physically, due to the nature of HOSC's mission, access to the system
for the purpose of creat ing a val idat ion test-bed was not possible or
f e a s i b l e . Consequently, at the time of this wr i t ing , all val idat ion
of the model has been done empirically, based on the unders tanding of
t h e s y s t e m h a r d w a r e p r e s e n t e d i n t h e e a r l i e r s e c t i o n s o f th i s
document.
SECTION 5
HOSC System Analysis and Conclusion
This section examines some of the interrelationships that affect
the performance of the HOSC BYPERchannel High Speed Local Network
(HSLN) and draws some conclusions about the operating environment of
HOSC. HOSC interrelationships will be examined by comparing figures
derived experimentally from the model described in the preceding
sections with certain analytical upper bounds which will be described
in the following paragraphs.
5.1 Performance Bounds
In Section 3, several metrics were briefly described as being
valid indicators for describing the performance of a network. These/
measures were labelled D,S, and U, representing the delay, D, that
occurs between the time a frame is ready for transmission and the
actual time of its successful transmission, the channel throughput,
S, and the channel utilization, U. In order for the experimental
figures to provide a meaningful picture of the network performance,
some standard is needed against which they may be compared. The
development of any such set of standard metrics must be predicated on
the knowledge of the network factors that affect the performance of
81
the HSLN. In p a r t i c u l a r , the f a c t o r s t h a t i n d e p e n d e n t of the
attached devices are of prime importance. The chief factors are :
Bandwidth
Propagation delay
Number of b i t s per typical frame
Local network protocol
. Offered load
. Number of s tat ions [STALL84],
The f i rs t three of these f a c t o r s can be used to c a l c u l a t e a loca l
n e t w o r k p a r a m e t e r known as a.. This parameter may be used to deduce
t h e m a x i m u m c h a n n e l u t i l i z a t i o n , U , a n d t h e m a x i m u m c h a n n e l
throughput, S.
Another important parameter of the channel is defined to be the
bandwid th , B, of the med ium m u l t i p l i e d by the distance, d, of the
communication path. By d i v i d i n g th is q u a n t i t y by the p r o p a g a t i o n
v e l o c i t y , V, a component known as the bit length of the channel can
be calculated.
Bit Length = —
Network System Corporation suggests that the propagat ion v e l o c i t y of
4 0 f e e t p e r m i c r o s e c o n d b e u s e d when c o n s i d e r i n g t h e c h a n n e l
performance. Using this f igure and assuming a t r u n k l e n g t h of 1000
f e e t , the bit l e n g t h of a theore t ica l IIYPERchannel trunk wi th a 50
MHz bandwidth may be calculated to be 1900 bi ts .
82
The ratio of the bit length, L, to the length of a typical frame
forms the dimensionless component called a..
Thus,
bit length of trunk— frame length (bits)
or
a =- VL
Noticing that d/V is the propagation time on the medium, this
component may be replaced by the HYPERchannel Fized Delay. Recalling
from Section 2 that the Fized Delay may be calculated as
Fized Delay = length of trunk (ft?
(usec)
the parameter a. may be calculated as
a. = - * (Fized Delay)
Thus, for a HYPERchannel trunk of length 1000 feet. a. is determined
to be .48.
An intuitive discussion from William Stalling's book Local
Networks, indicates that a. may be used to determine an upper bound on
the utilization of a local network. Stalling also indicates that
83
th is hypothesis appea r s to be v a l i d a t e d by experimental evidence,
therefore, a part of this discussion is presented here.
. . .Consider a pe r fec t ly e f f ic ien t access mechanismthat allows only one transmission at a t i m e . Assoon as one t r a n s m i s s i o n is ove r , ano ther nodeb e g i n s t r a n s m i t t i n g . F u r t h e r m o r e , t h et r a n s m i s s i o n is pure data—no overhead bi ts . . .What is the maximum p o s s i b l e u t i l i z a t i o n of then e t w o r k ? I t can be exp re s sed as the r a t i o oftotal throughput of the system to the c a p a c i t y orbandwidth:
„ _ throughputB
L/(propaga t ion * t ransmission t ime)B
L/ (D/V + L/B)B
1 + a
So, utilization varies inversely with a. [STALL84].
An example of the use of .a to predict the theoretical bounds of
a UYPERchannel trunk is summarized in Table 5.1.
84
TABLE 5.1. THEORETICAL BOUNDS ON A HYPERCHANNEL TRUNKLENGTH 1000 FEET
Fixed Delay = 38 \isec
Bit Length = 190 bits
a. = .48
U = —±— = .681 + a
S = U * B = 33.8 Mbits/sec
85
5.2 Correlation wi th Simulation Resul ts
The d i s c u s s i o n on performance bounds in the proceeding section
c u l m i n a t e s w i t h t h e c a l c u l a t i o n o f a t h r o u g h p u t r a t e f o r a
HYPERchannel t runk . This p r e d i c t i o n does not, however, take into
accoun t t h e p ro toco l under w h i c h t h e t r u n k t r a n s m i s s i o n s a r e
c o n d u c t e d , and t h e r e f o r e , represents only a theoretical throughput
rate. Nevertheless, based on the emper ica l wor th of this f i g u r e
given by Stalling, [Stall84], several useful conclusions may be drawn
by rela t ing these f igures to the f igures obtained from the simulation
program.
In Table 5.3, the c o n f i g u r a t i o n p a r a m e t e r s for a m u l t i p l e of
s i m u l a t i o n runs . The i n t en t of these runs is to demonstrate the
p e r f o r m a n c e of the HOSC s y s t e m o v e r a v a r i e t y of d a t a s o u r c e
g e n e r a t i o n r a t e s , inc lud ing some c o n d i t i o n s o f v e r y h igh Trunk
Act iv i ty times. Figure 5.1 and Table 5-2 i l lus t ra tes the HOSC system
model and ident i f ies the various adapters and devices .
Table 5 .4 t a b u l a t e s the s t a t i s t i c a l r e s u l t s for the v a r i o u s
s i m u l a t i o n runs. It may be observed that the trunk ac t iv i ty is very
high for run 5. In fac t the trunk is ve ry c l o s e to s a t u r a t i o n for
this case. The data source generat ion rates are very large for run 5
and are not expected to occur for an a p p r e c i a b l e l e n g t h of t i m e in
practice.
T h e r e s u l t s c l e a r l y d e m o n s t r a t e t h e r o b u s t n a t u r e o f
HYPERchannel under heavy load condi t ions . As has been pointed out in
earlier sections, the HYPERchannel m o d i f i e s i ts p ro toco l f r o m CSMA
86
-*02
o*Q
en
OS
HPU
Q
CSI
at:Uit~PHQ
OA
PT
ER
]
<
a
^
X"
r-rN«
^
^
ft 4^^* 5T
<rc5<n f
^ rH
9-
CN^?>
s
^
•'i rH>< r-H
< CN
CN
*\
-«CJ CNHCtl^^-'C rHH >~H
CN
4n?
CN
ro
i-HCXI
rHrH
W "— 11—4 CO
H f*1
fcj _:
> < <N
CO "^
O P-r- l—l
fcj 0 ^r- ' E-i
K U rHC HH CN
2 CN— m
0 -<u <-i
C-JCN
•
C
•H
CO
—ot
VUt
r^Ouc3
P5
O C O O O O O O O § '
^ 3CO
rH
c o c o o o o o o S
-5 ^ 35cy
o o o o o o o o o ^•H4_lrH3
o o o o o o o o o "^rH
00 ^c ^ O O C T i O O O O O . j j
h
r - O O O C ? 0 0 0 0 ^
O O O O O O O O O
rH rHo o o o o o o o o
• -H O O O O O O O O O
i — I c s i i — i
87
TABLE 5-2CUMMD1ATIVE PROBABIITY OF TRANSMISSION MATRIX
DEVICE TO DEVICE
110 211 221 311 321 331 411 421 431111211221311321331411421
00000000
0000
0.010.01
00
000
0.020.02
000
0.7000
0.021.00
000
1.0000
1.001.00
000
1.0000
1.001.00
000
1.0000
1.001.00
000
1.000
1.001.001.001.00
00
1.001.001.001.001.001.00
00
431 0 0 0 0 0 0 0,0
88
<n
VI
i
CO
O. ft
o o •<»• o o «n« n ^ v e r > i • <s o o o
U3izoI—IH
a
wuOS§CO
(ft
as
BS
o cs o o <nc4«n • «<S
•ft en i-(
r-» O O w>vo • « • r- o o o^ OS CS t»> 00
O O <A• ' 1 — 0 0 0
tS f*> 00
o o m• • (S O O O
S3l
COtaeCO
cdNM
W5
asMu.u,§8H^
Ma
^1zj«l
f»iz9
tsZ
S
O O OO\ O Tf
o o -«r <«
C S O O O^OO O O
o o o
0\OOr- t i - I C 4 O O O\ O O O O
04 <S <S
iin U o
a -Ed HO
co «as ««
zo
uMu,
1
1
CO
I
O O O O O O O O Om o o o o o o o o
O O O O O O O O Oi n o o o o o o o o
v > m « n < n f n f n c S ( NCO C\ t-t »-(
O O O O O O O O Oi / I O O O O O O O O
O O O O O O O O OV % O O O O O O O O
ica
uuZ COo ^MCO _1CO 5t-l >S OSco 2
M&z u5 Bat aH a
O >H< r t < o 8 8 8
«-( O O »H
O O O O O 8 8 8
^O O t*} ^O ^H f*}^ m f » < o r » f n 8 8 8O tS f> <S r-1 CO
O«•>cs
8 8 8
ooooooooomoooooooo t- O<?\ O O vo (S Oc o i o O ' r t ' H O a 8 8
as3
cd
8ua
CO
oMCOI—I
_]s
89
COz *5 T>04 S
^ oz a «O o u
H «»u to
CO ** >>w
S IM •»<O >a. -H
S 52
-J CO
B gCO MCO CO
82CO
co
CO
oSO
oo00
ONON
OS OSos os
00Ot
en
oOS00
00 VI
O <n •* soO so00 •»
I•nca co
to o o
CO »
57.5»«HCOU « -453 -• «
J3 *•J a o< H HM O MH O Mco co a•H '* ' •**
I I
CO
oMCO
Ct. CO
°s=tt CO
oo
<H OS00 SOOS T<s <n
M% * c>» so ooO 00 M f-l i-l
»H r-
•< iesO
O <-l
so r- os•» oo 01O so r-<s c< tn
g|
90
conten t ion protocol to a s l o t t e d p r i o r i t y pro tocol as the trunk
activity grows.
To demonstrate that the theoretical bounds from textbook theory
are not easily applied to practical systems the data in Table 5-5 is
presented. As may be observed the theoretical bounds are unrealist ic
and are never approached by the simulation model. It should be no ted
t h a t t h e s i m u l a t i o n m o d e l h a s b e e n c o r r e l a t e d aga ins t a f e w
stat is t ical measurements and does seem to be quite a valid model.
91
TABLE 5-5COMPARISONS OF THEORETICAL PARAMETERS S AND U VERSUS MEASURED RESULTS
S = Throughput = Total Successful Bits B = 50 Mbps Bandwidth. U = Utilization = S/BTime System On
SITUATION
THEORETICAL JIAXIMUMS
RUN 1
RUN 2
RUN 3
RUN 4
RUN 5
S( AVERAGE)
33.8 Mbps
.2978 Mbps
.420 Mbps
.529 Mbps
.761 Mpbs
1.293 Mbps
U( AVERAGE)
.68%
.59%
.85%
1.06%
1.53%
2.58%
% TIME TRUNK ACTIVE
25.05%
27 . 84%
51.12%
67.16%
93.18%
92
6.0 REFERENCES AND BIBLIOGRAPHY
ABRA77 Abramson, N. 'The Throughput of Packet BroadcastingChannels.' IEEE Transactions on Communications. January,1977.
BURK79 Burke, R. G. 'Eliminating Conflicts on a Contention Channel.'Proceedings. Fourth Local Computer Network Conference. 1979.
CERF74 Cerf, V. G., and Kah, R. E. 'A Protocol for Packet NetworkIntercommunication.' IEEE Transactions on Communications.May, 1974.
CERF78 Cerf, V. G., and Kirstein, P. T. 'Issues in Packet-NetworkInterconnection. ' Proceedings of the IEEE, November, 1978.
CLAR78 Clark. D. D.; Progran, K. T.; and Reed, D. P. 'AnIntroduction to Local Area Networks.' Proceedings of theIEEE. November, 1978.
CHU77 Chu, W. C. 'A Hierarchical Routing and Flow Control Policy(HRFC) for Packet Switched Networks.' University ofCalifornia, Los Angeles, Computer Science Department, August,1977.
DEC80 Digital Equipment Corporation. VAX Hardware Handbook.Maynard. MA: 1980.
DEC81 Digital Equipment Corporation. VAX Architecture. Maynard, MA:1981.
DEC82 Digital Equipment Corporation. VAX Software Handbook.Maynard, MA: 1981-82.
DOLL72 Dixon, R. D. 'Multiplexing and Concentration.' Proceedingsof the IEEE. November, 1972.
DONN79 Donnelly, J. E., and Yeh, J. W. 'Interaction betweenProtocol Levels in a Prioritized CSMA Broadcast Network.'Computer Networks. March, 1979.
FARB73 Farber, D. J., et al. 'The Distributed Computing System.'Proceedings of COMPCQN 73. 1973.
FISH73 Fishman, G. S. Concepts and Methods in Discrete Event DigitalSimulation. New York: John Wiley Sons, Inc., 1973.
FRAN80 FRANTA, W. R., and Bilodeau. M. B. 'Analysis of a PrioritizedCSMA Protocol Based on Staggered Delays.' Acta Informatica.June, 1980.
GRAY80 Graybeal, W. T., and Pooch, U. W. Simulation: Principlesand Methods. Boston, MA: Little, Brown and Co., 1980.
93
HALL78 Hall, R. C., and Widman, D. H. ' Implementat ion of aD i s t r i bu t ed Compu t ing G a t e w a y (DCG) a t S a n d i a N a t i o n a lL a b o r a t o r i e s . ' Sandia N a t i o n a l L a b o r a t o r i e s U n l i m i t e dRelease SAND82-0490, June, 1982.
KANA79 Kanakia, H., and Thomasian, A. 'A Comparative Study ofAccess Control Mechanisms in Loop N e t w o r k s . ' P roceed ings .Fourth Conference on Local Computer Networks. 1979.
KATK81 Katkin. R. D., and Sprung, J. G. 'Simulating a Cable BusN e t w o r k in a M u l t i c o m p u t e r and Large-Sca le A p p l i c a t i o nEnvironment . ' P r o c e e d i n g s . S i x t h C o n f e r e n c £_£n_L££ a_ j^Computer Networks. 1981.
KLEI75 Kleinrock. L. and Simon, S. L. 'Packet Switching in aM u l t i a c c e s s Broadcas t Channe l : P e r f o r m a n c e Evaluat ion. 'IEEE Transactions on Communications. April 1975.
MCMU82 McMullen, F. Telephone Conversation. Digi ta l EquipmentCorporation. New Orleans, LA: September 16, 1982.
METC76 Me tca l f e , R. M., and Boggs, D. R. 'Ethernet: PacketSwitching for Local N e t w o r k s . ' C o m m u n i c a t i o n s of the ACM,July, 1976.
NASA82 Nat ional Aeronautics and Space Administrat ion, George C.M a r s h a l l S p a c e F l i g h t C e n t e r . ' S h u t t l e P rog ram: HOSCOperations Procedures Document . ' SPMC 1.2.1, Revision 6, May,1982.
NSC79 Network Systems Corporation. A400 Adapte r Reference Manual .42990008. Brooklyn Park, MN: 1979.
/
NSCSOa Network Systems Corporation. Nucleus Adapter ReferenceManual. 4290003. Brooklyn Park, JIN: 1980.
NSC81 Network Systems Corporation. PI 13/14 Peripheral InterfaceReference Manual. 42990015. Brooklyn Park, MN: 1981.
PEC81 Perkin Elmer Corporation. 'Model 3240 Processors,' ProductInformation Sheet. Ocean Port. NJ: 1981.
STAL84 Stallings, W. Local Networks. New York: Macmillan Publish-ing Company, 1984.
SHOC80 Shoch, J. F., and Hupp, J. A. 'Measured Performance of anEthernet Local Computer Network.' Communications of the ACM.December, 1982.
THOR83 Thornton, J. E.. and Christensen, G. S. 'HYPERChannelNetwork Links.' Computer. September, 1983.
WILK82 Wilkinson, W. Telephone Conversation. Network SystemsCorporation. Huntsville, AL: August 12, 1982.
94
ZIEG76 Zeigler, B. P. Theory of Modelling and Simulation. New York:John Wiley Sons, Inc., 1976.
95
APPENDIX I
Summary of ISO-OSI Model of Distributed Networks
1. Physical
2. Data Link
3. Network
4. Transport
5. Session
6. Presentation
7. Application
Concerned with transmission of unstructured bits t r e a m o v e r p h y s i c a l l i nk ; i n v o l v e s suchp a r a m e t e r s a s s i g n a l v o l t a g e s w i n g a n d b i tdu ra t ion ; deals with the mechanical, e lectr ical ,and p rocedura l c h a r a c t e r i s t i c s to e s t a b l i s h ,m a i n t a i n , a n d d e a c t i v a t e t h e physica l l ink.(RS-232-C, RS-499, X.21)
Converts an unreliable transmission channel intoa reliable one; sends blocks of da ta ( f r a m e s )w i t h checksum; uses error de t ec t i on and frameacknowledgement (HDLC, SDLC, BiSync)
Transmits packets of data through a network;packets may be independent (datagram) or traversea p r e e s t a b l i s h e d n e t w o r k connec t ion ( v i r t u a lcircuit) ; responsible for routing and c o n g e s t i o ncontrol. (X.25, layer 3)
Provides reliable, transparent t ransfer of datab e t w e e n end po in t s ; p r o v i d e s end-to-end errorrecovery and flow control.
Provides means of establishing, managing, andterminating connection ( s e s s ion ) be tween two onda t a to provide a s t a n d a r d i z e d application twoprocesses; may p rov ide c h e c k p o i n t and r e s t a r tservice, quarantine service.
Performs generally useful t ransformat ions ond a t a t o p r o v i d e a s t a n d a r d i z e d a p p l i c a t i o ninterface and to p rov ide common c o m m u n i c a t i o n ss e r v i c e s ; e x a m p l e s : e n c r y p t i o n , t e x tcompression, reformatting.
Provides services to the users of the OSIenvironment; examples: transaction server , f i l etransfer protocol, network management .
This summary taken from the book, Local Networks, An Introduction, byWill iam Stallilngs [Stall84].
96
APPENDIX II
Simulation Code of HOSC HYPERchanne1 Network
> % % • • • • •
97
o*-f."In**!
t '
»J
r-
^-
I.
•
r -r-
•»•»
-D
1
n«—( .-.•*..,jr•v.£•
-3
-~
*
«
«
« b.u
* «->C
« 0-<_>
«*rf
« L.
O« a
a« 3
ti«
U*« C
•• C* jy <r~
b>«-<
* O C£ b.
* 4-. Q,
•_ —
* 1.1 C."
* —* c0— X
* c— «-
* r: > o^ (/> L.
* u «-• CC- C >
•* t. 3 •«—;.. ~ ^
* >- z,r Of
* x c,-I* +-> 4->
« C c•*-*• "~ **
* C C ^ t~ ^•^ ' j
•* r-^ 3 "-cO C rj c.t-
* -^- *' !. fi•-* O -r- ••
+ -!!• • trt l-• •— •* "Ci IA O
« i 3 *• r ••- A".*J u C C vi 0
* Y "~ -^ A tn >-».fc . *- O ••» o
* v *• ~i ; C"*'
i, •-* E•*• (b 1— • +* x w> d1 (/It- 01 3
T3 < 1 X'*- U V>
c > e « w(BfcO*^ C 3 CJDO m ao Oft; ** <n vi * "</i • O *fl l» *^T) TT ^* *- X b. (CjO &. ul— . L.
rs. TT •*• r C*tsi''- (b Xl/) t- LIX *- QT) (0 0 O3T. "^ O *Q QJ 3
u — O *rf C*t. j o a» w o a -*\r-~ "> v *- c — •
u a; -^«) C.T7 m U C1 >
f' jt t1tf ^ y*- w *-* *-j
— ro C 3 Cb, 3 -^3
j^ a/ wi ai o xl_T7 _rf U £
O C w-«- •-» ui tn3 r r- «- c *«- U V> I O b.C C*E— O— Cr: c c-**-«-»
•^ a c. — *" c
0 r r- u '3^^ t. i o -ji
<T *?; *' L. «-•*- 4^
n.-*- o «•- — • c O.TT; c c-
• 3 s.1 U C> -,'*l-
»J C C u^O t'— ' OC' *^-I— 0 C'"-^
tfl (^ -f~ t" -T O^ X V- «- C. U O"j in C O ? ••- 1C C . w t/> CiC C CJ '-/t Xt/lT c •"* r. c - ~ •
^_ •*• C I t»-^ b.U 4-J — ' "J
• • C^O • Oj — ' •- . ; *-Jt- LJ o-'jr c CX" Co r. •»- o *--•" ^ t* c• -: >- r o c u u.(_'
;, ~ c a r o -f f *-*- „. ; «_•
i" w I". O " 0'-*- i-C ^ C *- — ."—*' O<•* >• - o r. "? >— •" .-; r
•O-D C «*- C O3 m.f- «o «
^ in Oi <fl «41 1 Oi u k.4- C. l. c 3 «<» oc ie cjiE UO^ E •- *•^ .^ k ^ <^
* * t i ( f O O C «U1£' X^- Oa> **— » •- u u •
^ Oi H,
on- o c ac «C OU 0
^ ^ u c/ «-CT^TT ffl *. c *t T 3 a- 1/1 -•ifl O— i X 3 *C — i u •- a- cI- C O — «ii— ••<• C- l/>Oi (L IT *- 4.i. c o «- D a.
C C1 Cl u .IT *t1 C— ' C.-f **£> £ ~ *J — «l-o - C C3 t^ «~'P- •»- «CT-" *^ O
•p- Qj— ' O IT *
C f. C C *C * • ^J'*" T" •*"
•f-«»- L. l^ i «^ *
O tt w-f-*rf C '^ 3 T" *
— •• '— t" fcj 11
3 *: a w c C *(/I *-i£ C i- 0
c. > ". r. *b. </l t- T1
^ *- y. •- *yi Lf, • b. cC *- C T' C — •• *
t. > « ' ' : f Q *LI L >
C *- x irt n o *~ 1 L*. U t.
**- ."- "" ") •'- *
*w j*~ ~~ +* Ci — •"• - • • i c * . *
_» ~3 l.-«- O'*"
3 r 7, *-. w- IT '+r u L-r ? t
— j--»" • j- O *(."*•*-".' '»• .1
*
f^f^f^f* «« «« 4
« *« *« ••
* *« «
« «
* «« «
* *« «
*« * *
*^ *4- * t-
3 *c. « ^X •€
3 « C<i *
* >— «» *
C, * K— * «— * <rkk «C. * C.•^ « *L. * * <:
in * * _JO + *f -» t -
*•* i• *
•- * r2 •c. *
«— «3 «o « <
I .-*- *3 « ••r. •*C * •"*
>— *^^ *~t «r «
-»• «
0 * *
; * ** *
_ • *^- « «
,-s
fl^e
•*-,
oin'<3'-tv r i
ftJ
— ' 'C _y
O C b.i-O C
o 3T i *->C f O•r** C
ma«- c3.^ -~i -•
0 C -:li"- C
3j. G. t-c*- ~3 r
f1" *" 1^S W -~-
*-••*- £•
3C"- • ••-- C•" ^ , _;
C_ r 3 KC" t. -
t. r >•- :t' t_ r
• *^. {.' <-- *»-• *-r . i/i
I' 1 -O < C1. C 1' i
ti-r- . . \ ,.t-. t. _j • v-; _•i, r, r— .- ,•• •*-(_' *- W r- '_ 1 .* U
*- C. i" - k-3 1" • *. , ; t *. ' «. 3
o "-.* . ~ ;•-•*. *-» « c > :> --r tr r : . : n t. :r 3 u f - -• i r <-•. • • : r .: c . • u
c xu «
U I--*i O C
•-— 3
«- I- Cre *
u/ t_—'3 O
• - C Oo j-
Cir w»-*i "- cf >
^ C 1- C -I O L —
•«-— e.— 2 ^
c— «> ^- -.'i. >••- -:t i^ C F c.
6 E C C -
; w t- •-• '.
c: I I I I I-1 I I i I l
•U' ~.C~CT>_; — (--• r*lf .—
i i i I I i i i i i i i i i i t r i t i i i t i i i i i i i t i t i i i i ' i i ' i t i i i i i i iI t I I I I I I I I I I I I I I I I I I I 1 1 I I I I t I 1 I I I > I I I ) 1 1 I I I t I 1 I I I 1
i i i i i i i I i > i t
OO C7O OL~"> *D ""»" •"!* "ir~> ~*C* ™ ""*'"~.T*'r>T1 CTC~O -^ ~ ™ ."t'^'" ~>T?Tl'T"» O:3'~iT "~I1—1*'~ ~C" "'T »"~1"^Tc~ f~-r~ r r*-"*r •!"-•." •'"•• ' * — i.-r- T' •.—•c'oc— -ic ""•*"" • r*t™ r;f • ',"•'." f -'~ r r "*•„"*'* t ~ T; ',"".'' c ^" '"*"•v. •»(•••"• i,"1 z~ T _j: "* r1 "" l> • - - - - - - i T -)-•, _; -^ ->^_- -71 -i.~i 'T! - —. "• ~~?: ~- TI: ""~ ~ ,~ '»-" "i "* r—""•• ~ ~ ~:
" r ; r r
i % « « « « ^ « * « « « * 9 « • • •
98
•*•"« C XW^O m**- C
>.•---«> 3
,-> to in
X E 3 C ~ "3 — — <»- C C
C— 130'CO--C*- ?7>^ w*-
a •m o
C >O O fti
O. > b>-^ *» o
O> O 6o u -^
"O *.M- 4t w c
a,-*»-^ C *» i* "* t r t b . a j C
a.«/» 3 coacv*-O <rt **-t
C *-C 4-
-r-c«- 3 O 4-1
crC— • C"- O
C, C«. -,c
T* — • L) L.
OV C_u »-* c a
• Cj 4, - £ Qj <D •» *•*• C1 Q. Q.> U.ii t/> ^> O O 'Q "50, C L C C C C'-*- CT-T
O133E t " - * - O X X C r t 3 T
•*- m «- — ~ x— <— ' re a- t-C"-vJ «- C— re r: E £
r-, .* re * t »-•>••
"1 re ^* *••'•"••*'••' ^ ^^
• "*0* V "*V •»•»
13^ ^ • J/c • •» si ft, ** •-*— *Qj •* _ ' O fc. C1 1^ • » re C
Xi ; x rec-fcCjC-^ca,n n Q; O*-* re- c f-b. ••«* Q; .. a, C C Ob.» • C. X*-">-->-f;- •• ••fO «£ fl •• iTJ C C4-
r x x— •- n c• • O +* fJ t- *- »- o/-^'-—A/ t_'^ _»£ ..0*J t »-* • -
T" •*•• *- C C" QJ b^ •*• — • f— ••••• -^b. > */)•#- T- j - j j c o O X f O C »*•>C C— 1- C— »' 3 U*' ^ o •% •»
Qj C nb. 1- ^
(_.-
<4 -
C
_*•r-
i*
(•
X
L
^-
•»^ -3ac
i— *o_^• r-
Lk.
CJ>
"•
•u•Tb.•i/•C
P ,•~>
11
(K[_X
<f-«•
7;O
^«ui«,c
»co
• f-l
•r-
—
Cu
_
"
CL.
c;••»-
f.'
•**cc
•LJ_^-\J
c—1
c &c.-*- cc
L,*s-• a*
c a—*- n n
••:*. LL,
t".<+- •*-o o
I:-— ir tw >
b. L. a,o oojj *- eC C 33 re cJ" T >—
< •v F^j
• *r'~b. K »
t re >
t »o*- • cC «•— 3•-*1— j C
K -
• • X •t" r •u b.«-•V b.k-1
*J "Td x.-; •• *-- b. b.«; t, b.* «-.- .—t . a: . < * • •3 "C k-
T) *•*• i/> u i/i (Q (Qt c a> a**fc- b.
•»«J w'. (T> H (L C"'
— — — C C O
0 •- O C.L.C- G •- — ••-
» -T^- C,
K 3- r'c t.f~ih-**£_ • '•3f9
^ £ i'Ci C*" 'C ITc * c •*-*-.• E re3 a. c c c +~ x:b. » •- -^ b. • <T" O O'
I- <fl C O— CC »-*-. »c* o rer~ •»3 a nai t- b.C' Or c •• c uc. b-X C< fc b. U •(•". «io *- 3 r.jf •• c «c >
• -tC. b. L » L. • • u-
• o — •* -^ «t—• w ft •>•-» b» C 0- X»"
£, L C IP b.ci U *~ji x*- » ^ x»- w x r& c; ->•-<,- •-*. t r. i.*-< l_ » ». .1 b. A
C b. X - ^ r; . %? u -;P- rj — »-i C -J b. 4, it « ,
— * r *• c .^I...-- ••, ' 4, •^1U" ••
«/> J. H- X LT Q' C <C •'.• 2.-^ «^ X— ' •r- > «-^ ^ — -C l-f — r M- C o —3 C 'J •»- *-••-* •• •-: *-b c wfr £ C -u r* Cl . .
•>• .j x xt_ & < - • : • • c —, 0. (-• L.-r- L ^ 3— •• b.C"; LT •-''•-' ^- b. L '- M . —t: 3 * - f x r— r L -.3b.T-L_i~t->r- »*.*,».
"i
C
— .
w
,t_
•*O
r-.
, CC
—
uc
c^T-'.•—,
f -«v
C- •-TT t-O *•-.- -
. . »; _.' ^, 'r, 'f: *-• "• i* *.", i— C
*+j ; t t-: «. c ;* h r. : r z. t
I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I t I I I I I I I I I I I t I I I I I I I I I I I I I I I I I I I I I I I t I I I II I I I I I I I I I I I I I I I I I I I I I r I I I I I I I I I I I I I I I I I I r I I I i I I I I I I ! I I I I I I I I I I I I I I I I I I
_ ;- ;" c i • : : :' i r : ,•-•- . c~ < i ; . '
- ~ • • ~ ~ "
r r r r c c c c r r r :~r : r- r r r r r r- r r : c r : c t .- c c: c c; c- c~
. . . .
99
zo
* c *• «•• »— «• ** » - •* «•» ^- *
« *« t~ ** «* >- «• ••• w *« ** L . *
* «
« *
* «
• *
* «
• 1 *•
* ; *• ••» ^ *
* i- *• •* I •»* «
• r •• •- '. *- : *
*•
* *
* •—* *** >i-* c• •—** i« —• r_
« ~« •»-
« b.
• C
* ">
*
« 3
** ^-
«• C• ~* • w«-
- H: C
* H C
* — f« "3 V* "".'.
* *"*
* — L.« ~« f. •».
r -* * c.* -* c
CJC.fc •
>• L_J >
_!•-•«- *—
• H C- •H KC UJ
b. t— fc.i f—v.. O*•• *" * ri, 1.1 _•:;—• t;/ s, •»— >-o c z.- »ol_ 'r- «'r--C.
.-c o ~ T i-> r00 «- - -
c ~ t (-•» _'3 r c.- w _i
r -. *- * iC *' •£" T t " *•.; . »- r- .:•- l.' C t*C --•- 0 C ? ".*T f- C — u.C.
3 rr-t£ •» ** -— <r :•
X (- |-
"* c > r* i3 1, — • f w
— • T/, - i_ ;.
yi C — j",•* > •_/ ^C C ^ —*.' — • ^_> s.
CO\_,t^Jco.w.p.
fjCo
^>
c':
r i.*- c/
L-; j:
^- C O_l O
» — * - « _>— -i;»— N —
•-• — rt u
a, . o
^ .tT ""
c- •*• . .oO C t-
^ C Cc
«-r J U
z A u
• V/-,\
c3L.
L.0
1 *-
C" C.1r. cCC —
0 C*- ^
t3 Cc ct/> *--
• » fCtfl C • * •« t>
•» I' C. » *»c 3 •• M n • .^-^r •* •• •• c. •- c:b. w-O X C C
«»((—»-*- * "J* - :^
• • . j *• n c o*" \ ^oi xc o •_- «\ •» > n ;•
— ,--— .*•*-#- — C CT"" C»; r." .T n • - v. i, i) w x — ' — ' C — <.- * •*--*_....«.». o c — ".-
..- ...,-_-.,_ ..*, e*--F-.r- rr-r-
• \
*-s
\ri/iu
0-VI
C
*•
.f.+ ,
C3L.
L •%3 *^C- &•- Cx-^-
r*«* -T
'"
C C^ C— • — • c -^Q, C— C/
. . ^<r.
,
%
CenCC•fbl
I/I
•^
CinC03L.
**
3
•»•
HICuu3in
C3C •*
' " '
1 X
» -5
>
r *-• r— • c — •»c — t
— Q.^
,-x
\rL.
O«->ftl_a.caob.4*
C3
C
C
cVc(Db.
o»-
-J
t" •»« '
~-
c c^- r c— — ' ~ —•0. :•-- 1- r.
. . f, -r-
L. .,^%
• % i* C u.
0 —V £ *-
+ J U.
A
V ^ U•% \ O
^-r- X
\ ^» UJC » — >C dH —C t' r»
C i— «— <C *- J
«.-^ 0 H-
•0 Ci: —c r*~ou**- u. • >
r . o • » C-J«- ^ T.
i'aP 3 — t:lt\ >«• b. ."Z
t O * K b * -
^""lli. 1 L.
c c:r-\ — * - — * — * t
C fc. — • "
C tl •- *^ £ C"— *- "..• c: -- —.; <v i -^ -^- c c
•- f. t . t — "•
t.
• «N C
i-
* *,/•.
V •"\
r- '*-
r— *ta..
t*Ht>
C- *-
••- 0 C
CO.— *-3 l^r 4^ J »• i
^ i.1 — >
*"* fc- (t C *»
rr O b, —» - 4- C< f
0* C •<-;^-^», ^, . ,
t »: n n i,
*•'•-' i. c. C ~•»•- ^ 3 ;C •• 0 C I'. t;
t " * - *- ^ L.*' :> 3 IT ; 'f t. f. : . ^
c » - i — .:: t. b. r. -• —• u t -.*- —
-• t r c ••- c :• 1 C - * - t
• :•;• cTrc-r'r r r r ^ •; rr c": re ;"T: £ c c-c t ; c*:: c c c-r r r r c rr-r ~ cV r r r c'r re rr r !?'f r!Tc .-"j f r'r: £' c c < - ' • : : : .
-r -r .r >' f r • (-- *J ir -s »T r : "
" •*"c TT C r-r >- ^ »r »• r r r ^"r r -f"»rt-* c "^, •- j . • - .- • 7 »~ ."• •..- -' ."» r-
J
* •T ; 100
CJ>ctl
I Ii-C.• •c
U. I' r_ .
C C. t -
Q*UJc f
£ «- i3 (0 I
C 3O 3 OT - IT
C
• »0*
yi-»-
«t oC XO*'
3 O«.or*- o -on-E u <^ f
T: a. 3 c -— 012 3 -E —
u c 3 *J
!•.- _ ._ 01- O
\ i, u.»i~i. 3-~>
a
o_j
• «utC-
^c•"••
^_;
•.Ia
>** uu
LJ trtH" &l
*. » ^^3 -».UJ ftj*
t --•*-
i,i ak O —3"- »-
0.' U 11
0 *e
c.^^c »-
3I- OH- *C
C M
5 cZ 3
*-^--4i *.^ O— E
a~•fk. C
^1a
C ••
t-
*01T•« w
c a/
3 a
•ca,-»-— 0•^ cai;
>*•u. O
11
c.
c ••
r'
»- ot^-p-
>r> L.Q C.
a- c
•F--^
C - i-r c.^
•P-T-
r-« i- L.
u.. in ou o oa. c> ti,
r- c c
T; r c i-
c r-"- —
'^£n^
•>Hi
TT
C
^_•^>
•*-
''-
.
C
•-•
t.
"
b. .a> •u a.
> — t,
r- d*-
t r-i u^.
^ k.- c' Li. &• C.
C t f C.
L. C C«_ a _> wr-
T; c i C u.
• •c c— _
-.'r.f ir-
I-.C 3a
o o c.»-ti ac.cc.•^ —f — j O • •^•^•^••^•^ -^ «^ Cr^ k. u u k.
d ^ C.«-k* WJ f i t / >w l
i. u c c a. tv. <•» r. j-.
o v. i,] r. _ c c : C £
t- r. t T 3 C C O . C . O ,^^ w *^ ii «-;. - *^ •* w «^C C C •• ;,-.-•- — -r-
- t »i-«--;iL;j^w *- *- *^. •'• ••- L. 3
:: n .
r: i. •- *-» -»u u ;r ^
0 tn • • u. C
f. C
c c *---• •a. C-r:
*-^ •«j"- <•- X"
u •«»- *-t. U T •»-O C W 3
»- C.•*"*'rc * *C, *'
1
(*-
/-~
•C C-UJ
C^f
«.•- ».II <
— >
L i-
O-'-*- J
*
-* tC T"%
t"CC 3
- >^- * c;
i i i i i i i i t i t i i i i i i i t i i i r i i i t i i i i i t i i i t i i i i i i i i i i i i i i i i i i i i i i i i t i t i i i i t-- -j .•!i ' '. i I I I t I I I I I III III It II i II I I O I II I I I I I <- I 1 I I I ) I .*•- I -.* I 1^1 I I I I il I *? I I I t I I 'I II I
.' ' p " ".',"•; -" '-" '-' i" C" C C;C ';• C *~ r-t-r r- C • T "." ' "'"**—*- r F t—r r- •%;»*«" V1'* C"C" - « - « — » - » • f jr */ ^j u"' u .i.~»L~.«*r W t-- r<" ••" \» ."-?-•"_: "I" CT" -TO13T" OT^ "5CTC: T'l"4' ". .T^^.'r-^--- ^ -.*•-, K K" ^•~^•^•^J v t^T-J > ' J > >J^J -.?<-r-J^r-J^'.t-/ f O
. . . . . __ ^. , . , . - . . , -,-,-- ^-j-^-^ c r-^r r i- r : •- <r r - T : ( r^.C" :"T - . - • • • • r* •.?(-.-or '. O;~ r/c~',.-t~ <."',"•'"'.. ,_ „„ ,._„_.,.. .-J(- — -. _-3,-3.3-i.-i'DOTTi^^c-TTtoor^D o '"oor?,' - •*" •pv.r'oric* -;••• r-n'^("-r»n -^ooor .- ; : c :. ; - r : r r r c, c ; r r c c r r r r r r r i - -c c r.c:c r c C-LT r: c-^ : "t: c:: c : r f ," c.r c c c L c.c c:f
* ,-T- r-» r '^ -T w- VP f c ^»- r -^*"•-" L - -. » x -- ^- ,- ^- ,7 *- - r r ^ f- r- r i* ^ r -r- ir.r.r ,- f _r .-r T^C •" *• r r *• r -r ^ r r;r.r * r ; r
-cr r: c- C" r -r r~.«- *. »: r t r r »• r,- .-• r -<v c r C- C 2 C ^ J »" :• :. " ."."T r r i r - i o - r •." ;r .r jr...*. r.-rj* « .*"
101
ouJh-<
— ->• x
\ r••- c.
•-•- c.••—
w •!•(. U•- t3
•
L' I,'
— • 7;
• • * . *.— -
* -.
OX*- <Q
**•*• C "O— — O • *^ 4" f» C"— - C C 3--
^ (,- <e 3 -^ «- •*_, • •- L. — rf V> .v _J
o C *- O « «- OL. in r"3 o i'o i/iC. £ C
T c. T w e *-0. L. I C— ' .BA. *-<- , <C fc- "O ^*• i ;• ". r
.- .-* t- w C t/i o• • C" C ••- w 3 ^
•*-•*- u<"- C 4- 0, u.t- c o t'X; *:*-•" -j * 1,1 c 1•T - t"*- f. OT ..i: *' c *- o u c— f -„ C t- - t 3 3 «c c <- LJ c f
C — C C- f • f-
'•••'- l • c u E o cj <
— •-to >^C"~ <C,- - ,_.._ t-i.^ _- — c
~ ' ~ > i u w u r — ••^ L- i_x C 0-"- C"- •• ••w - *J C t- > Z Xr i . t L . ^ c r ; _ i
<•_ — c* J- • — "* i c r *-u T •- — i* ^--7*- i
V ~ w c* C O —w "- - X 3* •*••*•• C > —-.:-; b u v C c r. ; .cc* »- cc.- i- 3 o*- ^ <-•
inT
^OUQJ
(—.WJ{*•u: {•*— Z2C. T'fc
<*—r f t
*< C.
I • •
O k.
cI/ trk. a
c ct.
t* >— '.-c t•^ >k- a.r: —
u— • &
*I ^nO •-
— —0
-* _t -^ a.
!Z " -j< c: «t£ 1 0 ' * * -0 Ct • « OC* C- *~ *—C-
t > oi." c. ** t*. ^>j •-« TJ "c *^ ^r»-« H- * «— «— • :K— «•; V*T~ t/i * «. ^ ~*<r _( a. c c ;/_j 3 O u •• O u ^j c tr I
/ r s : - c * - - » " C t - a . * L f -K ^ _ O 3 ^ ^ _ 3 - - I H - (
^tJ — » t/lu\ ~ l t l C *-v_* V
u> w> u—p t w. i— *.- « • x «- r ^ _» » t a * C " t - a / c - > * • e x . ^
r- u- n ft; C *- * C 4 * C + C. *-* *••* ^- •*-•* O f C- O C 3 T- C. «- C •• ••*£ - CC- *-i - •• «o».r»-a. * ^ ro»^^ »-t3 i . - - - . jwv - * O Tt \ . c' .»-Tf-* *^ *- •* T*n *- o vc. 4 - r c cc
C- \ t E r x / U - i C f E-^w.C C •* «*k C^'C. L;^- •* "T- «^L •«
c **** «-^ * ^ i.-c<LC."C.^ ^ c a - c . " c c » - e^ ' - a x. c.«-»* t f ;•- •• ~' — •
*- 3 3 3 3 3 t - 3 0 C. *."- 3 C C. J Or- 3 •• .••«- ^ u -~ C » t -• • o c. O.Q r. o- n *-*- flj w u rfc —•»- ^ *- c i- o £ c.O 0 > * » 3 3 3 3 2 — '3 r - l /<^C 7 «-«*.* C Oc. T; c occ. c: c •* r •• *-t •• •»
O > ws^*^<w>^> w u £ *- *-• :i ^: •*-• % » : — c c c c c c ..->*.••- c .1-1 *-•••-
«T *•* 3 u c. c a. a e v i- i- . c i c -- *- Ci • C> J.T-*-w^4-^ .- ? •' C
i. (i .— ao< - i - c i . t _ c o i_ o• •* •• *-x~. r ? r " r ^ T " i ~ r ' * T ~•1. -J II •* C
C "3 w» •
> ^
I •" I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I. I I I I -.• II ' I t ' ' < I t I I I I 1 I t I t I I I I I I I I I I - I I I r • t .' I | | ( | | I ( I I I I I I I I I I 1 I I I I • '. I I I I I * I f ." I t I I I I I I I I
: - ; • --.-,----,.— ,,--,-., -1_,.JC _ ,_. ._, ,_. , . . , --- ,.^'.- -; - .-- • -- — -j — . ." ) .- ; .-. ? ~ ^-v- | T --
' " : . : r c -. r-r r r:c : T r •_ c~c~r c r ; . r; c c T t ^ r ^ ~r " r ^ c r - c r r - c ' < : -~- r C ' C - r i
102
o c
L,
6
oao^._^-<^^
t£
^j
C-IS
.*._
(•*-
w
•*r,
c
...u.J3
C
c
-
-,
cor*^_i _
^UJr .^_i.
UJ
* .
^
«-;
~
*~
i
l_
^T
c
lflw«-
enC
V.
a--^
X
t.01»_,dC
•*_O
cc
aV.0
i) "3X
O ••£_ **~ f, \
— C > *-."i" "• c
3 u C C
- C e o —/ .-. & &, L.
-j C*"
**
**
**
**
*•*
*•*
•**
**
');
r N
ET
WO
RK
'):
AD
AP
TF
R
OY
A
DA
PT
ER
');
rl-
:')
;.-
iTru
nls
: 2
, ')
');
C. C 3i 3^:
21 *. x0 fr -=•— i*j c:
c^- c -
p-ltj o vi yiI." ^.'r-L~>- C
cic; 31-L. a,
L : .^ j*\LJ C^ H- 3> 0 Cl_ •*O t. E *-s
a, 3 ui
c— c3 x 3C C <-
-3 0•- EO* 3
v s % > ;^ •«*-*' •» * "* -xC C C C C C'^-'C
0 ; f t , oaO ' <t i -- c
%
X"5
a«TJ
c.
L-
L.
O
••li
(/)3
w-
4.,
a ^•*- \
C C v
— X • • «l/l VI r—s-*-* I"1
C C •••-
3 3 •— •*• JL. U X
*-»— w
*t * ^~
c c -.-a.3 ^-1
i: ~ 3^^«j ^ CCT'C K- 3
o* >•-J»— '
% C *-'^ li — Cc •• c Q-— •
u O
jfU02«.«
o.-—
c•^
V)L.
o/4M
aIT •»^ >tf Ul
L.
^ sa
L. rga.-— <F. •»-3O; c\ 3
;;
"T*C C ^^
— — C
*- I. C
k.
a•a
c3
COt
•* e^* o* u
» eo
O • L.
•c r. •»•
y, •• »-k. * -. «-a>*- • c •»G. » —'-•" ^ ,- *"••" c «r«i" O fc. -w E» T- 4; T . C O
Er- C. ».»-L_3 — • rj \ t^ •_>; • t . \; •»- (.1
t- < % r.—
4-,*- -V^- . %w
c c c c ^ ~«— .O -^ -- — ' -* C
"t, C! C 4. w v *^
': * ' " " * -——*- "
^ O k. k. k. C It
O
a
*M » C II
-> o <t cuj a, o
* ti L. a* o o
U ;.• « ^% -^ u.\ x
— > O ~\ a- A• • • * • c. ~ v -~X *O v* u. \m i c » o-f-* u-* a> o " c*-*a, r~ — * ^ c_
£l (/• W>, * -^- •• 3 i. r -•- h. » " * - _ - » —•r- *f> O T t.l_ C Wi * »' — L'O C" QJ O * v'-- u-•f xf x r *- G,U "«» W • »C w ^C r - i 1 f ~ t '^ '- 'Z'^-
r— <J -^ »-* 'U w- •- *.- C. C = -^ t.^f ^ >i/ '^i*.'1 ' '^.^- •%*- j u i ••• c. TC.> C1 "C •*" » ^ * - k w J * ^ J '0 XC > » ti L.\ C.. •x- t o.- — • ai "••".•..,.*-"—-_! *^ r- • • t - * j c » * ' :* CL>JC "3 r1 • r , >j c..
r. c T- T r- .» -- .. i.^-»X3 — -• 3- .. c» -^ J
r.*-i- > -J * - " 3*X T- r - , .
• • t . X " C C r - \ * C * - ,ton C — — . ^ ^ u ^^ >
•^ .. -^ Cl u C C r ..- ".- •vex »- .- .. — — — *- -. *--. _
t c ^ - e c ^: •• c t.-it-.-— "• .'. " . .~rf*»C<->-^ Ol • ^ ^ ^ / d k . W b. M. — . -
t t— T; c. t -i f — .. < ; r ; < • • - —•-•-•cc.*- a,—
u i- o-— i- >;
I I I I I * | -.. | r- | C' I I I I 1 1 I 1 I I I I I t I I I I I I I I I I I I I " I I I I I I I I I t I I I I I I I 1 I I I I 1 I I I I I I I I ' - I: ' - * I I I I I I I I | 1 /_- I | | - | • f ' t | I | 1 | I | | | | | | | | i--. I I I I I I 1 I I I ' ' I I I I I t I I I I I I I I 1 I '• I -" I I I I ' - *
" c : c ~c c rc-c'-.r c r c rVr-Tf-r c" !•;• ~r "c. CPC"C : ii'rc:"- c'V r : i : i r r c~ ;'c c-: r c-c : re re c r -c r cc r r c r ^ d •.' r : : - ^ ;
" ^*L H- r r1 c r* »• r. ^• xjt- v- r -*r T »-*• ** ^-'w *" " c ,c C »-• r r- «j L. v r ;• c
• t t f * t • • t • t t
103
n.cw
^jr-j(iV
.
(J...
>?\
^_.
"•
'.-^
— ' t:* —>
..- .-t_ ^,
diy
t e
s/
s> :
*> ;
ev
ice
*f
J :
1 )
;
C-
0.— cC -f
k.V4,
•— • k.
•* 3t- 0 Oc • *•- ^
C. C r:
c r
— > tC •» 'i *- u• *- t -- C 4.
-: i *- ii v; i. c
, > w * V c-C 1 t' '* r? C
C ^- *- t. C —— • ~ *•— — • c
-• :,* c --
•-* ~ ;.• ; •«- ^- '. •*• C i- i.r c.- > - _ ; i.
• »
*o
t
*-o •»•o «- ..
< *ri C* O •i.j k. ^Cr U.U *- O*k. (A U3 ••- k.0 ^ 3
° U^ r* S
tn — > >— <TJ
u C 4* OL. U *- C,3-«- a. *c > ^
»• t *; L0 •
-. t . ,^L. r- * V
^ _ c •— •n c c c^-
— Cu. o> a —
.. '_ O-r- .^ Q
.= a •<- k. L. o.v •- », ti 1. 1» c.
L. jf-
^>
0 *
vt« •» a*
-* * x
* • tn«-•*» c
C"; w^ *^
«— ^ *J Qj •*X N CO
4 fc{,~ >^>\ -* V) «
a* Cj L. a-
C ' 3 L. ^ (/•O O « 3 •*.*-c" t" c C'-*- a*r- * O C;* Q,*-
u - C t, C i.- c*'L
* £ *- 4- C k.2 •• c c c .. a i;
r:^ c •»-.: * r-» a* I' * 3 (b
. . . Vfc. »W •<-
Ii C C ** 4* Cw t -
3 c c, — <r o— ' o
= >.-. Q C*i- W+-z w k. t- oo iua (3i . 0 t. U C.- C.
orL.
f-l "
-1
4*U•r-
>Cc.•
t.(Lf
O•j;
^
II
a-
f.
o<».h-
^ oc
\ •*-k_t •*•
u ZU3
to QJ
U 0;C
a > ' D C*- a- ofl fcrt ua- wiC a. ^&( O x^U U
n >•^»
b C jf toN 0; C »••" £ ^ C
i i. QJ K- ^*£ 0. •. t- 3 » ^•*>•- •« w U
•»- f- > 1- >- » »•
• C »- *- H ••C c w •• •••- ••
1: II »- ^ \ * *
1- 3«- » > »C -^ ^ ^ k—~ • «^* ^^*•> > 1' (^ fr Cre u b. •»- c — '— ' —c * - * - r ~ c c t a
t>- -• t 31- C--— •-
1- l^^ « C i.-^*^ 3••- — c *- — '.1- 4, 7- . - f
urc
etL
D
do
L : 1
, * : ' ) ;
t R
es
ptl
l <
> '
';
re
ad
ln;
pC
1]
= 'y
')
the
n
0 -•• U!b v \ «^ O>
O • C CCI j) t j< 3 ««^1 C • •
ink— i 3^ ^ ttlX tu l.r-1 o 4/ <fl
3— » ^ ^ u ?5i- > C. > >- u.
•• c> V 0. « II II-:- . a: •• ••t r" « i IIJ •-» % «^ r > nr u' "J- ! i iCO; G- l_- —^ »• ~ U C.C C
c c f r: 3•a c n n t . t- >-
II <-r- C. -- C C C• « r*-^ c c~ t' c'
_it-^-. — r? •- c.
L 3
^
f-«
^*
V
l/l;
r.=>&»-u.^.
t,~
(^>»-•^
-"^k. Oc **•
«-• a
—
T" C
r-
TC" • 'C T"
CC
1Cc
c.rj
f-/--i T-
•T)
;; w* ^<u t. w C
o*-- c eT" *" ccr r-C •-- r C
^£ L, .•^ O Ii
w •j'*- — t3
• *<* «-c
C • V-r"»-
cx
eo
o*
ua.
a, n
S<C c\C EQj &•
h-
u- Ca.
*^— ii
ev C
»--:i *." /.L t Vi
•^ O k. V—
«- k. C c.
1: *- C Ci/i t* r; u
^'Z^1"r^"*Hr«- c c
*— i 3 •TJ «- I ' rre- -i . «- C- -
-I*- t-
» C 4-
i ' i i i t i i i r i i i i i i i i i i i i i i i i i i i i i i i i i i i i t i i i i i i i i i i i t i i c~ i i i i i i c ff -*t w» i i t i i i i i *i i lit ii i ' i i i i i r - i i i i i i i i t i i i i i i f t i i i i t t i i i i i ' i i i ii i i i~ i „. i i i i i i i i i t i i tit i i i i i i »
" ' ~ ~ •' f "' r '• " tr -r r •' r • * _ - . • . . * • - > T ir T r " r ;f •" :f--,i(^;r :r r -r r f f .r T -r :r r f r •" ' T T .* '- T r> r f r r r :r •• -r T :r •*" "" T .• f c " ' ~ ." f r . -.• - • • • . * - - . • ) - : - :-.-'ri-.*TM — . -i ^-, • -»• • .. • -• . j T r o - T o )— -c~ ~ r — 3ac • r*- "t< : '--.-i.'~ -•>--• .'^ -• •: roo^jiTr, "•-i^o-:;T:>Mj.r'""''~. —-- ; . - - n •/ -:- :• - r : ; ; v r r r „• c c c - -c- c r. CT. : c: re. c'crc c r r r r -:~r-cc r : ^ :T-C- r rr c c .' r c~r c e r e . CT ••:•: rr rr c-; ;-r~r r-
t- • • r • T "*• P :K *.*».•»,- f. r1 f- r •• r ^•'O u- <. f c 'Cr ' t r - r v.-L-<:r e : c r " * - - P r - oi"-"* c - r* •-• ."o^1 »* LT y; r
• • • • • • t « « « « f t t » « « *
Qi •«. Cc ~ a.
• *c*-> &Of£
C ' "•- •"» w ** •»*« * fc r-»^O L \ I / » W ( / 1 » » ^ - ^ O C 3 O « - T— u a- 3 -^ 0. *n •*• T w; ffl*-> UN • «. O O £ - E = * - ^ > " s . Q C CC* a *•* i. C </". • 3 *C ft* ft1 i/i trt t. *••- c-i— t a» 4. 3 " O » ^ r ' - * ' * c r ' » - 4 i qj ^. ^i- -- «— • - * - O l-i •-> >r*.^ ,C*+- • N C C : > v
(? TO t*! tfl -» t i c — , -. t 3 n -^ 3 »f. c. *- 4. fc...* u. f - -r c: f. •-• c *. > u •
_• tCC C *- r '.- r t- u— »- »~ ^ *- i: O > it — — ••- — — »' » r > e * * - \ \ l^ ,*-i. *- G fc..*-«- o -i ^ *- t.1 ill •O'^'*-1D'^1*^ a O 3 u- w 3 Q.» c C.T: w i. k.. f- "w. t. **t. * » * - . * * - & «; »_ w c- w. n: • r « - * - ~ •*•(-' n
\ 3 O fj *- G — > Q. * G C O> k_J •« Cr C • u . . < . £ f l U ' U C C L. T • r
L- F c r c » *^ ^ r. L v» 4» '^a* •*• •*- v r t; -^ c a <^ t> a. f c- a» * > > n *•• r • f ~z r c •- • c i
c r" L. . • -.' — — c T: T r- *»-r- n -- 3 *• u u t- u o. £*••• *•" i "-' *- t. *- *- •* c i r~^c r ^ r.-'\ f- < C .— . - ; • • - »-. Cr- (C *- (-1 VI w O C I r tT - . ; U C C " C C O CC' C * -^ 1- O'^ *-- C C C •«<-'• *- •'
-•* •—- *»- -u u i. L. •> ^j i. wL-T-'t- t- c c- a> re C k. a- - ^c^ - - -— -^-; c a-»— ^- H- .jr..ex *^ c oc ^ * ^" " •
— i_ —i. — t f-i •*-•»- T c o c- tf1^— it n C. ~ *- i-K- K *- •• (.•:(/» f; r 3 *n -T; c- »-• ti^- x i-x. t. 3**- c- c -• •-•"— •- 3 •*- 3 -• t. i- n c •- «- o i) a." "O c c ', r. -c c ^: c— v i.-.- c o •»- *- —• ±3 L^ j *- -^ •- t z: i-;- u- « - » - . . — d *-c y c — —-^ c c — ^ *-«:r.c c C.C.T.. >• "- c : ••• f. c----
"- — — o •- c a t i* ••; c *j c n: T —• t t- t o L • Q c _j f. _j <r <- i
Oi- o t. f_ *- *- t " " C C Ut- *' C fc'C- -. C. C-."- O <~- "•• C 7.•-. u .ir- ,.•)«- ^c c c;a - •-ji t a. c*- -* "- c_^o.
c. ^ 4-1; ».:. it ** •- -j v i. c— c _• *
--•r f « t. -*..-; 3, . » _ •- c i o•. -~ i » * c *
i ' i i i i i i i i I t i i i ! i i I i i i i i i t r I i I i I i i i I i I i i i i i i i I t i i i t i i I i i i i t t t t i i t i I r i i i i ( ; • •i i • . i > * t i ' i i ^ i -.• i i i i i i i ^ i -< r t i i i i i i r • i i i i i i t i t i i i t JM i i i i i i i i i i i i i i i i .• •. i < i i i i i i i
-. -.r r?-^C' ^r *. •:"-. fro^nT.^'^--* — onnc . " ~- -.?T -. .p j— •-• i r^cr-rrc-c oo-^'^-oor^-.r^. D r»c"-- D~»C-"'- "c'c.'. : ; c ; r .- t ; c.-c. C'-r c- r c_ • c r c~c r r r. r CT-C-C - c ; -Tcr- re c- c - c ~ c ~ c r c~c r r L" : rr c r c r r
" r c~» r « . f w < T - c ' C r r r^'r.. ^ - •. j- -j- r~ r**' r c c ™
C . C ' C % * * * f t f e 4 f r O ' & % t i l t t t ^ , % , * 4 , u , w /
105
••o t;1; o•t w
^- •-
C •• VE
a c~ E •
t r
C >. •Z C I.
*- J- "»• c.
c. *c.
I- t •<"•*-^r- i»
T"' T T
r C *'• i'r-c c <-*- i.wk. k. c. •: c
' — c c r. • •
. " -«:»- -.." ; K
-" - '"" - U
-_< t r. ft ->." - " - ' " "3 V 't 7?_ k -,^_ - - _ . ^
c *•- - • t: i. : •k * ' C *. t, • C. »•
^1 *•* k. — ~"l t. 1i .-: - * - c t .r . "7^ *~ *U »* r^— t: ,•? ; "~ •*
— ... ».. f ~ ._- * •»
•- " o ; c
f**.
V
^^_Ul
i-"
L J
a:r-( ,-tc-T
*t J
L.
- 3
u uc
-
C
1
C
•QJ
tr
c.
x&>c^o
QC»*
o"f-3c.
1X
«ctC;
.C
C.• N
0. *-. "' 3Q C.
~"Z
i C
1 Cr
i- U>
T: a
o
c.
t
*CTIQQ\
»
C*o
• v «•x-* **-3 ^
M »3 (Ji•- C»-- .C TJ
— - ^)O QJ•^ X•r-k. » I-S
^ C.?~ w •
• *o (>r*— "C •» t
\ c.* ^ ^»--! 2?
«- -^ D U
«- 3 X T— ™ c> 3 *- a-U C . M •* «t *r o oC. _J * * O 3 ^* • j O
*^ C Lii M j* it *-»^ K ^
'-«-' ^C »- *- k. *- *^ -^ #-
•i* U U " u C. k. U ' »
C f _J»- ^s-r. j*-
i." cV,
• *
*«41«*
k *••» *V 41C *ib *
*a. «^ *
c. *«»-— - *
0—*• « »- *c c «o o~«
*^--^ 3 «*-•<- C.*C kl*- #
•*" o *C UO*W U 1 *tn trt> «lr 4> ^ «C-^< **- *C> ft* C *C.C— *
c -^ «
C C 3*
0; t- X *i* --1 C *3 *T *
U *O *W *
o. *
«•-o
0),-•
a *^~** Mt/»Oi, X
••" ZS•»- *t
IS> f-rg ftj
_^t'^
a, «-* »ck. C **a* o 3C JJ C.c *-m c- 3.ii-* O
k. 0r- *«••-
«^.^0^ C*«~u-*- x
•{Jtl <i-
^O
C-
**« «*.* +•ft
** •** •*«« «<•* •«
r «« «** +«* *** •*4r
-• **
« **
•« •«*
•y.* L. >.-. t/<x * r. * r i r .
»•« »« » ••^ * *- *-• *--3 « 3 •» 3 3
c1;. « ;.' c, c
•" 3« 3 3 3t; <f + « ' , « « . . <tt, •-* * «^ *^ ^*t. c * c • c c:
C « C; * C- « &
»•
Oaa3
<•''
U1
CO
^ •*-U f
o <uJ k~ t,
i- C
- c.f -• :.C — .-c —"! > «.-*: wi uu *-> •:.IT C >U.. 3t_ „:>- —= ij
^ S-1 *• -
,.-\ .. •. •- 1, o „-> «• ,r. u-.'-i.- i - •<>• r. ( ^ o r cr. (•• r_ r . -cr r ,,. r - ,
c i - k. •
« k .-^ »^ » » • »•» Mb. » ^ - .
*» •• *»^ »' ••* ^ ^ - * < ^ . £ i v * ' i i » ' > ; * -3u 523 . . 3 3 -r 3 r -"• r.- r
•C — _ 5 w ^ '_ . .- C c J .. .-.i
3 -.3. -3 J 3 3- 3 C : 3— = w :;».^-*i «• ' ^- «* » ' 7 - « . T * « Z » . : •^* ^# ^/ ^/ ^^ ^>i^»^ ^^ ^^ ^. *
c c c r c c r c r ^ . '
e - « t t » £ < • - o - « c c « t ; . .
c~ i r j i i i i i i i f i i i f / i r t i i i ii < i i i t i i i o i i i t i t i i i i i i i i
• r- : r r . •(.;. •_"*" ' " -r , ,-r r r-r ; ir'*-r-r-*-
~ ^ -c -." >.:ps -r -r-p>.^-r.
106
Of_*:
•—_*1
r <c £
.Q t_>9,
I- oiN •*•
X **3- •«-
0 0^-. •*•
L.
t z»
M (->^-
L
-*: C*t- TT
CJ 3
«_C. O:: c
•t~rj *-
^ C
^ ITci.
— c^ co --j *-
o
r r . r r t - c r . r .
*.
^ .;
- 3 r « - r r > 3 C 3- : -' • _• C Z C
J J J s J 3 C 3 t 3
L « C.i ^, !;*• C5 °CJC ^ L. C C C
E
*- Xin *-
1 X 'fG yi u
••- *irt 1* d
+• inO "5
tn -*
T] x 'w»c. -o n:
0. w 0> tl *-t'T3 ^ ^
^-tl >»• tA•U C 3I
<s* c— ' U f
^ •»• *-•c •-0 1/1 Jc •- o
L. -rfC Q^ — 'L. *J «^
T u3 *^ C#' ^ •'*- tc j- ':
£— a —.-; C ••-u •— cfl
C r .ur . [
' ' :' s• : -
o T r r 3jr.- i C C C.
C - 1 3 C. 3
"li~Ct;iSr c c
« c c. « c
V)
11uk3oift11
•£o
0«C
Ik£
^
>.O
V)Co
**•i*ftk.-.,V.
cnw
»,
c.
-
«. 31>C
•- 3
•~^C
a
« * *
ia3
V)
C0
*rf•0
a/QO
a<
1.>(nu
C3
^tr
\\kC.«rfc
•-•t,J^r-
fc-u.
CT: '
-, <-
* c1- 31- 3TC —
3 I. 3^ »• C tt
c r« c c »
It«»(QE
•*" •
1/1
o>c11c1>u>oua&^
i/>a>^3O
. >•*-
.„
4^—t3i/)a.u
r ci -
0
3 r 3c — c
3 L 3£ '~,
C C
t I
1c0 Cu O
Q* U)
_,I- O0 u
IK•V Tn 30 —-^ u
C•^ -^I/C O
rc «^JT nu
v\-c ~&f — ' 'I- 3a, in
•*•- t.0
(L.
c. f^ a-*- ^-
cr . Lf -
L^
3 tn
r-O C- 33C- k-o.
u 3 r 3
w^insc c
* C Ci
4 .p. « .r-
1*UL.
3OW1
a/L.
^
Q
J
a>c
u<*Jo
•*.c
v>+*cEa.*-4-(Q
Cof-t/1
• »
{Jv>cr;L.
C "
u
^ 3-c.
t 3
t-JSL
* C
*-*
c<Q
ft"UCISEUQ
f
a
a-4>.
(T.
X
(rt
O
4H>
rgU
•»-•C
•r-
C
*-•
C.V)
3
a.k.T)
or,
E
•-
3fcC,
C 3
^SC
C.
* •«-
• •I/Ic «0
•~ «*•^ 4u3 «o>•^ «~o
c *0
^(0 4-^3 <E
• - *u*
*11«-•
«C
•r- «
in *CO •«
.' «•*•T- ^Co «u
4
"<o ...C U~. '.' . u ^ LA-- r , r . r * r -
>
.-, - - 3 « 3•»- ' .' <J .'
H 3 3 3 3
Si. i i * —C C C « C
« t C » C « 6.
.
.
r
^
CcQ
r
c
• * c _
"'-J•» o
•- cu u
]';-
o
^c.
3
;• 3 3 « 3 3 3 r r = roo»o ooo.'. o o
t? x c « - « - * « » . - r- t.- ,>r\ r-
>3<iC c . cc c : c c t ; c c«- o r
li C >- ••
1 f 1 1 ^ = 3 Z 3 3 3 3 3 1 J 3 ; _ ' C
i i i i i i I i i i I i I I I i I I I I I I I I I t I I I I I I I I I I I I i I I I I i I i . I I I I i i r I ( i I i i I I I I I I I I i i I I I i IJ I I t | t I I | | | | | | | t I I I I I I \ I I I I I I I I I I I I I I I I I I I ( I ' I 1 I I t I I I I I I ) ' I r- '. I t I I I I I I I I I I I
r-f r ^,^-^-
- ,• ? j->:- T—-3 — - > ~ r. :;.3>-3'.T: -j-r ~ -oo-_-.TJO j-i-o r- • •> -> ~c.c r ; r r c ~ t r-c m-crrc": r r rr c c~i :"T r r i . r r ; "
> -i,- tji-i.Tcr-.C'-T^-'?. .30 "•-•.">''••-:•• n ~TC- -i j-.-r r r c ; r c i r re c r r-: c : r- r c • c : •..-.:
•' • >-•*•" ^: r 7 c c" r-c.f »-v- -.• f»-* - • •- »•• r- • - « — « — . - . r. r.,rjr -r r,-r
107
•C --n -a —r-
>- a.- ..t; ..— >- cI- - "Co— t ••
o•c 4, v>
(J,^ \
I *
V
i. t/1c.
u ,1.&
<•-*tflc -rz r 'i4.*- ••
i'-ii/* *—3
•* X. ' •* OL"t: o»-4 *"^ *V
»—'* •* • \ o-. **• r . *-
\ »(T3r r> •• r cu ui * f- '• u.L f> | Q.3 0 1 • • -
_J \ \ >•-w C
•* »— ,c +^ v w «-»3 *•*•' 3 »• 3
/w
rvj»-
•»»• ^\
* *e•*- ui*- <L•0 wre uO 3-J O
in o* T
v ni#- m
• » o v:&l ~ L. .
L. O•r- *»- •* U«»• O*-* k.
f. 3
13 k. Ci^ 4- • • Otv
O ^J 1/1 "O •-j e a n\ 3 w o~^
;* v. c t-* \ 3 o c*
r; c t. of>'; »VJ 3*^
r- »• c >• • ^ C o 1>
k e •-£."3 O •
\ \ " f_r~•» =^(
*^h «- «-r- i.3 u -3 CO.
er
',
Ce
wfa
un
: 4
:'•• b
yte
s)
;
E •• •3 fa.-C <M
•»- ••
a* i/i OGu
<p- C >'t> QJ-I-•^ C *- T'-•*J Q «*-
t-<\ 3 1^C' *^
» « % - » -r^w vnt*" f" 1 " t.
•.r Cj*• *. r>^<*-
X t •*
•• 3* » * r- - • 4 ••»
+- *-, j *^r-; 3 *3
3 3 ~ 0 3 **<r «; u * «•' o
t</a
no
•• C'":
**•
\ ' » Qj
li *-* *-\ (Q
& oro .. Ci_ a,, c,3 »-OO TJ •
liT t f"» V
Ct .»o a
-.T ••- U
fl W b.r; 3
.. 4_ O« t-t"
cX Q.. *
(•**-
» 3
* * "?_* '
mis
sio
n
inte
rva
</>c<i} /*->L.
** tft•», QJ«k: "» uC» *.3 U 3L. & O
t— >/i in
\ «-- O
*J »Ki^r e
3• « »* c
% rO
\ * * *- r**_ri
•• > *w— «>
*- t-rin ;•3*- "U.O C* -*C~
3 K «- ^~-c.t- U O*-
r*>
^N<MUJt-J
•—,U'
C)
•^«. r1^<w
C **• —-__1 ~
C. k-J-:• C ^
~~- ~ ~r: — i_ 3^CO£ —_l3 - 3 - 3 T. C C. ^ 0 « « — C
c_> » - : > . - - « ; w - ^ -_ I33 —H» 3 « 3 x -r- C. C «•' « C C e/ • •
OC 3»- 03*' 3 ~x a> «. u x <u u «--. * u3d ^-J 3»~-J ^ ^><t O C ^- C C
t C i. •• C. C. •• t •• i C.
c. c c «; r «.< c c c•- *:- j c t c - ' - a - i - * c cc •• ^ii!«
I' •
L ^ ^ U U ^ U W W^ O
t. ••£ tr-i-— —••c •- ^: i> u o u i
C/"-
r: _ii i, v TS L. r. -i r r.-. ~C •!-
c •«c '.
I I ' ' I I I I I I I I I I I I I I I I I I I I I I I I I I « I L I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I o l.~ - •' '"- I I '' ' 'I I I I I I I I I I I I I I I I I I I I 1 r-t | ^ I | | | | t.^ | | | | | I v-i ( 1 I I I I I I I I I I t I I I I O I I I I I I I t I I I I I I I I I I I I ! I ' I •
• _- —^---•-•-*-«-'*»-r ;r;rjp
?:c~c~c-Tc fi^r c ; :"••.• f : rr cr-t c c - - c c i c ' <. : : ."; f c:r n .' ri c :~c ; ; r-c~r-c c •" r '• c c r c r c:c~r r
? •».•*- »-f r-r" - •••-••r-t-t— r-»~«-«— «-*rr r -r :pj"
- . • . • : t - -. ; . :. -
J J
I I I
108
f
~* r;
c«- o •• r
- v • •"
0. C. • i. 1Go- C w «- C- ZT • i_f T; i- • c
c r •• o 3" J c 1-1T f • O "^ T T r - ! C ^ O O * ^ K-
- » i i - i O TIC D.
*- t' Ox, »- 0.' O -J E I- CLc ^ r c u r 3 c . i ! >-.tJ
a > .-•• c > ; — -J". • — '«i-Lj
-• - i- - c c - T: «.. o c - o 3 o ct • : o • j^ — »- <d •-*:•- *-•**- c »- o •+• o t—
> r. - c » c«--i--; * o « «; u G c.
> ;~" c c L. ••• ' r, . . ;; c o _•• -7 " C o - c
;. _ - : c r. _ c e ••* C-—T: a; •• i.. . » T rv * . » « - » . « •- i- . 4J ^ , r- 4_ C t - C * - ' -.3
- , _ - » - * - ^ _j i. i;^-^- o.- .. — «..„ r ..- ^._ i.
^ - t. 0 .*
c-
i i i i i i i i i i t t i i i i i i t i i t r i t »-f It i i i i i ' ' i I i i I 1 i I .- I i t .-. i i r i i I I
r-r • • *.*fc~ v r r~ r r~ »-r r-- *• - j-f f.C rtr r r~ oi." \r:r p~cr - r. r^.p- r^p ^ p^ ^; ;• ^ j- o~ c~ ,v i: c'. 5- r «."• c* r r r r r- j- r
VI
c c
De
s c
r ip
t io
r u
nk t
x •
3
f 3
t. tfl
1- 3w yi
«- CT w
~'Z71 O
.-«;L. t-i
C »'
1. k.
a
i 1 1 ii 1 1 i
JCIO^-T-c r-c~ c r
r c r r c
^ « > ..* r**- ^
o.o •• rou.n
UJ 7
t -i ^ 6, " "*C^ ^ £^ ^
CJ>— L. •*
«-T O C. <Jz i. n •• c
>^ 0 ut > 3"•CD •-> 3 10 t O
L. 00 -i ^3
*- O O U ** *•*C f ^ F: C «-^ O t- U ••- 0 C
> i' EL 3 C > 3 ^ -0* > 3 > C C- O
• a • \ * % <• t. n k. *\ • » » -
u. r *- r 3 3 r >-
c: k. c.' c c c o c - c .c — > r; *-«j*^*-»-» **-*
fT. -f '. T C- • .' k, fc- (* k. *- U • *.k, r: *" r ? r* rt r* r*. T' r r fc -y c
1 1 1 1 1 1 1 1 1 t 1 1 t 1 1 1 CMt i i ! c-. i i i i i i t i i i r i i
r r r c r c c o i c c ' c c - c c " c c c c c
VIC
"re
^e
sc
rip
t io
f u n fc
I*.
> )
3*-Oi -—
k V.
u a.
u-3k lA
C 0
^"*-* to
X *•'
f 'W
C i/>
k (,->
C
^
u0
r.
1 1 1 1
1 1 1 1
r- r f r- r
jrjr jr-.-r :r-C* C C C--C
0; •*
•-. c *->rv- M *^^^ % Q, Cl
a/ r »•- t0 '* C "a f i •* •«. c. i » • • » - •
o c1 fc-s e* *-»• »- C' *.'•••- f
4> C> •» » •* c. *i • • C — •
f- T' » •*- IT C. fc. k; ,'> O < ^ r : " ^ - . . w
» •»- *~v > c. k x > . ? * • > ;*- TTuJ -»-y-N »- 3 .3 t y . • »
a. •* c -:;c u- r t- . i
O O Tt-'S I 1 •• *-• E r- K o o k . c- « • k
Q F C C C k.-»- K c* kt > *_ •-
•C Z.-C7; 3 3 c- c> "l— >»c > 3 ^ ;• c r ••: '- ^. - ; •
* . » c - » i "> t * * * » * * * " * *
3 r *^L ouy r^r' O C J - J * — . P L - _ >. ^ .
£j .-r- ,. r - .-•
c a i c- c r c c -J ~ r -. .
k ^ c k c. c c* 1 " *" c «' •
k » . r * ^ * - - a t e r : ' r - r ; ~ - : '•3 -i*
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 t1 1 1 1 *" I 1 1 1 1 1 1 1 1 1 1 1 t 1 1
rrc r CT c r c'-cr i r r-c r ,- c : ;-;• :
c- 1 c c - c c ? o c r r c c c c r j c : ; ;
109
r
j..
1
'.-
• -">if-
•;u.
C(/i0L.0
k.
3
t
L. •> Cu C__ •--
f.••*•
b.—*- — •*• cll 0t
i- C«~17, *-•- L.
c ic >
•X "
•- J—
'- *.^ 1"
.^l_ t"
;.-L, l-
co
^
b. J • * • %0, _J ^ — *
c, w •»
C 1 £H II
... (D «* • i" r.; ~c "> Of *!. <t •- >- ctl . * fc.1 • •
c c. *- r * ~ \ ^r- C 3_— f f
> •«. t- i/it- i. •* b. •%i. O > "• I-*--w'~- • •£•«•-••o t--*- o > < c ~»f • o «-KI a*>• *. T •!- c c r; u c. o
> " . f— b. .; •• U 13 •• b.
— "- > L-J > 3-. > 3«- ~;j'.lj •'-'•> _J 3 ~ v O C O Ou ••. f T.C- r w r.L? «..(.-,c b. *^ C • * O **- >•- 1— f : c
~T! »••[ T- L7 *"-•( b* * •• ••^ ^' «— i. «— t^* i_ » » * * ^sC ""r- ~. C L C* V > \
- ( — trc i,- r ^. b. »--*-* o C"^ c D c ••- a; o a» 4. ti <cU F L t" C (T-r- fc U U > U Ll t-
j _' t: w £ . ' t- /--~ c. w- b,TT I' t. i.* 3 3 ~ C > 3-*~ > 3«i >3> ; T C ^ ^r a* o u c^ o ii*-• <L r. o > 3 > * u. r L-. o -c 11 c
u «--•* - b. f- b. - - .» • . • * . r. »f.f j
r r .- i: ». L. u r = r ct r» r - ^ J — »_t_Ll • u
C: i: »- i- c. r 3 3 h-
c , C M c;i; r c c c r r c c c r '
t- t. u. t, r. c a.- ', C' o *• a a. c-3 C > A ; 3>c »-*-*-*-*-*-*- »-•- -^
tal •»- t >;").:. .* ', *T b. b, 1- L. b L U b. b. • *
i- ^- '•> > > - t f £ r z> — "; r r* r* "* r* *'T O C> ^- a
^
0
*rf
•*-b.uy>'i*
0ub.3 •0 Ci- C
•^• • tflk-"->— '»_i— •c. c
u•>
L. <*•-«- C«~t
t" **»•• ^
- (V. >*- cfO~ cV. O
*-o«-
_J -T
3"c •-
0,'
b. U
3 C.
Cfuo,.
b . ( / ) • « • «0; t/1 ^s *•»O C ^ • J*"1
C uJf"- *- *-
^ < •• <9 (0• • orjx T?? WT f *t »f
a> LJ >• c.c. • % :c • •
O O IF* ^* ^* «~ f 1 — f f
> •* *— h~ b. > % b. • %(L c> • * ; . : * -QV^.»O/^»«ii ^-^ ' > > w C «~*fO ^^'f'l &K- T" r • •*- »•' C C U C tj
> "3 c * > b > r Q < * b . t c < * b .•• -»-^ > C. L.t > 3T' > 3
*- "C i» •*-*-. t - i ^ f y t c t i Ot- •» C T. L 1 0 C. C- f. f.c o^c. • » ."!' t-'i e- »- cr c:T" *-c * ~ [ ' * ^ c . r ^ * * » ••<r ^* i—,.~"w- «o U. • • •• -C "T r- f C~ C.' l'» * * V
O C Tf ^ h-..*-
*-• <J C T J F - J lZ " *4 ;OwtC 'C. C f c ' E O r ~ " ~ F O U > U b * f —'" 3 3 ^ l_ 1. k- f -Vr- t- 'ir-r- b.
•C z: CL- 3 3 C-, C > 3— > 3 (.-:
<^ >3> i"*E^ t<; '^ Gv O o C O ^-
u *- .*- b.£_ b. r- -t • H-.r f^r »-
3 c ^ t o t - u s r - r r c .o K-* -h- (.. >c; r^c* c i . * « t « « » ' » " » » « ^
a. ci' a u c c c c c c c c c .-
i- u [ . t. r. i» c, i. c cj ^ c c • cS 3>- 5>C .•' »- ^w. . . -»-„.. ^-O C O t . T _ C C * L ""'r- -^ '^ -r- -^ *^ f -f~
t- * r"*' ^ c r. C-" ?- z~ r rv r r t- 1? r* T/•3 V C> .0 '•
•~\i~*
• *
"C •**O T"
• f C> ^-
«- c «cU— C 3
— C C
u n o ct/i .- if"**i( »< - •—
a a. eft, 1/1 fc-o X C" *'•b, *. -TO
3 L *-O «-»-*-
L- <L C**»*.£
• • *- X t*
b.-r- a,o r c-r*-• Ul w~ c — -^. f- . -«££ b. *J
f - C •-3 o
1 b. V "•
r; x i > •
» C. Vl u".
; ~* -TT C. t- cC -" o • ' f.
•^ c r *-
r '-.:•"T . ' U * *IK C I? C CU .": ', . ' Z
Ct.c:.
. "fcj%_
1 '' i ' < > ' i i i i i i i i i i i i i i i i i i i i i i i i c~ i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i r i i i i i i i i i t i i i i1 ' ' ' ' ' ' ' 1 ' I I I .V I f I t I I I I I I I I I I I I I I I I I I t I I I I | I I •. . | | | I ! I i I I I I I t I I I I I I t I * I I I I I I t t ' I I !
- \ r- .••.- r* -r i *.r •!*•!.*»£ r . 7 1—*~ i .t .r t '* 'L. •-•>„ Of •• -f ' *
" r.r»— «--."u" -r^rr r- »^ »- >c-r ' -r1 *>"• i" :? f -•— ^-«— r-p--'* r r r r ;* r r' i > *~ •' f ••' T-i' 'C^»- r- -vj <«" C" r- **• *' >- i." t w' i" C r *~T C" - '- ' ' '
. ' ' *"- ~r r - ^~ ,• . w' ^* ov o v -. v.'-r r r r r r ^. r r- r- ^ "*c~r~ ' ~ ' ~ i . ' ' i * ' f f * r r ^ ^ ^ * ^ - ^ - f r - ^ r r r ^ . r p . r r r r . r . " • f ^ '• i~ :
• '• • • ~ • ~ ' ' c. ^ r c .- r , : c. r c : : .: ;-c ~c. r ~ ~ r r c~r r r r r r : ir c r t ,-T : r c c r:c c r ; rc~r r-r i err r c-r .—r r r r : -o ,. :
: r r r ? r r c c c . : c r . : ^ r r c r : c c : r r c c c r r c i c r - r c c c c c o c r r — c " ~ C - C c r oocrr o.' c r c r CT C-CT r : : - ' .
J t O 9
110
3O O
"C V1
er- r-*u; -»L C 1— >
Q O &U £ U
L, *« -I—
3 >C x o
Of f - f „"^3 • *- •
f tir~1.1 1 i* •-•ifc-.L'i-
O u C< • k.T u u L. ai
o r-*- o «^T r . C >*- C.
~» •."> -V *- f£ ^sf L »• r, — Ti- -. c- • E «£ &;
• » 1 u r.r-f- u-• * •" •— 3 •— d t—• ~~ > i w - v • • »-
. J L. — •
c *.; f . c o x k. : -^ z • : •-•-•*— t <-i
r- * "V r ' »- Z. «^ •- ^
<"• -" «— IT x •-• «X1. L - *. - C "_ t — C«• T fc-C<ii*ir*fr h-
; c ..-;•• K> c r. °^ t".- x* * »- t *- — t_j» *~ C *- I L"". f— I' '' ~ ".
•»• c.• ' * ' » • I T
•- "" ta-4 *• L. 1
c ^• »
*" ^ T"C&• w
f^t^f^f
4.• % CT>
— C<C CC' «-c:
C• ••^
<^L. L.
c. aC"i
a. L^ wC C•— e• • o
•T*. T7
<- cI.: <",•.1 L.
ot. T;r: D> a.
•v trtd C.
JD rtS
•*- oc a." •-• •
<t*-c — vO — F
— u 3«-— Cu f^vcoc
»
• w v s- w w
•\
I *
n
-co£
* «
O•»• • *
^-«T._j^.* tjouaUJUJ
c^* u
<^ <** •••»*
• <-r^ ir rO--O "<-p f r .^- O
o***- r u n c i —
0 ^ '" C.' *" C
t» £ C
c• * otf> • * *^c^^ wO vt t/)
•p- C *-"•»- o c eQ— G, t/»
• f- *- U Ck. QL. k.u— 3 *"t/J k. O
Oi u "Uo i- a- a1
& C.C >^U £_*-•»-k. W — *3 U U. *Co «- o e*/) 31*- k.
0 O• •tik. Ci. C.& • • > Or
> i.— £.p. Qj 4^ 4W
ill *T3 W0 C 0 C(b '** k. C •c. n x
c- T; •••k — ' O k.
<TJ .C Wl **> •- c ra*^ -i £
. -- *- Xu c c *-• *J«^
c^ c *-•*
^ — —t" f ." .
3 •* C^.T2 ^ •-- O
o i.*_ *-• *";Ot.Q
C7;
<—E
•^ 37~
L. O
•* t
k. > V
Of •>-o a- ia/ u '-'4— 0, »«
c k. c T:«-» c^ c
T c -:•• *—t ' ^* *•••
t» a. .r- •«- O
^ C. •* X1 *-K ^-* C-
o r e• • f 3 V
ri *-* r*0.- ki > r»
«- T D C 4-.a, t? *- • * «—
• rf •* c — k. »• .JO— u C ,ii i_j *
Q <z o Q tr k.•c. a- u c a i
V ..C r* '.. t J.**". u .^»•*•• •. < c r ..r
*7C^ JJ II i.-* 0 C
» 5 £ — k. >jci-"»- c c1 t *--i r?1 a*> 'c
E3C
>t-
^i_C
«»-
E
—Ck.C, CX wC k.3 3C O
("c*U T'
— c> ,-u.
•c c-u
k.-*-c »> c
'*• 'Ct *C fL. *-
r.T .TC "JC r*r :: »-r-»— C
— u r
^-•w •'i.i
O~c
<-\
T:c3Ou~
«- co -3c:*^ **» •
t_ ',r cC. 3 r"••: o —
I. L
*-* .t"> »~ *_
V. 'O ••"
C.' C >* ' *" - ~» t.^ " • -'
c i- ••"-' T»- »- £C *^ fc. ;
• » r. v,- » ^
t" ;.~ i> r•^ V iT •C H * '-
t_ v -j - : .•«— IT
•01 * —^-~
- <-r- •• , •_•-
C T~ C I "• _ . .1
- Ll — f-^~
-.- o'. •„•; i \... S
i i i i i i i i i i i i r i i i i i i i i i> i ' i i i i i i t i i i i i i i t i t i i
i i i i c i i t i i i i i i i I i i i i i * i i i i i i i i r i t i i i i i i i i i i i i i i ii i i i i i f i i i i i t i i i i t r f i :• i i i * i i *-~ r i i i r i i ' i i i i i i i •- ' i
-r rr c r t- : : r c- rc- rc t cc r-rrc-r-r r c c c r c r r c r.rc~r c-: ; L . c c c r c - . c c . r c ".c c r
f "^ :T" "D "3t:. 71 D ~/m ;•?:': — - - •-c r*r CT irr-c-r r r rr re c r - c
r. *.'"''t," I,".' ^r ^T '.".'"" »ti-»~f f C C ~ t - r f w.^r CT c » r >^ V.-L-."r re rrrrr-c c-or r •• •" -•- • - - / " • " - " : " : - - " • • " • - " • • ' c'c cr rr-
»-' e" C r-»- r. c vr >.- .- r_ r c c « r
.rcrrr c c f r c c c:cr-;r-»-.-
V • V «
Ill
^r,_;~ • V
O r-i^ n
• » 4 - 0 -
O Ur~* C i-
c:. T o+1 C c/i— C *C r-iu- t/i c.
b. i— i* - • o xC ** oC C. •-
C >— -3 c,C nj £ic •
L. !™^ **•
v: 3 «:u C t-* weV t. L.U li 0.. >
i» •*: -_j-- o u»•»• i ~J iu- • * — • . c.. 3 r- -P-
.r n «r
*^- + . » t ••• %*— x
• » i. «— **! L. o"°~ t; t- »•<i i: c •• •% > i..
— C C, *»'* — u c w
Ol •%
r~ T;Cc.
CD *-C "•
•^* c ci- O t/1
t.-r- C3 l/l "3
"Zi
C —C O
c•1- *ri •^ .d«-
•»- *- e* a.— * t* 0-^— ' fc 4-
1 «^(J C> »" CU ^> (Q
• » *- ^TC /3 ^ f* Wr; L. fCb'»- C C. d>— 0 C+-
C C. trt-*- Orc u i/i E
C'*- tr tfl•• rs c r T:O w C <- O
•^- (TS*- ui/l O t. It
•»- »- O V".
— *- 3— C «-'*~O & o ^u; c >•«- •t C*^*^O
•f •-,
c i: ui c-^ w " r^•- • • 1_ U *-*»J ^ ^ <J-'~
C V ». C 33
UQC
L.
eh-
c *-O> * fl
0 *- X Cl
X — Or~ <TJ <B "•^ ^^— ' >— VI!-» C x C C ••"a C. r-tJ O — '
U^;«-.T; •*• — •3 C x • • -«. . 'O E -- — •*
OV) D i- -- CT3 -^ | . o —
r- > r- . xi . L.M-> il x*-> O C.4i L—C- »•- 1- r l>
O WC> »- *- L - < * Ct i- u /•» x a r- c
o 3-»- v a, *- x £T: n o> i: a n — **
^ t/14.'E »(0 *«^vi u ^r^ 2 L.T- TO ^L fij O «JT *•* <f C ~ «^ ^^a, u Fr- > *- o c ••:*- •*• D«-» Oc^-V •*- O ** I"a > r-i-c t CT»- c c
V C k- • »- •*» «i — ^*T; c. w o a. c w c— ••*- t. K-
• « "XO' -C*- 1 * -^* / 1 -p- — .^ C.** t-i0* H-Tr^4^ C *- / C* O — ' ~J uin O »•' *— ^-^*«i t, \_» O *•- C *-
•% —^ iT*i?t-J C "C E £* *- ^ "~ C L.u rj "3 *- O* M < r-^. oC i*. i C Q. U • • ••-ci •-• «-»o j:*- sC II C f- V «"»*r- C^ •• *-r- it *-> -F- ...c T- o «- 3 t.-• C t- U »C C 0 «
O •• "•*• _J• • -p- II ' *• _JV fcl ••-! •- -I- C
•*• "• "-1
-> — O s-c o u»-
» •*- tJ C •»L. »•• c « * * • • - T:• T O C> r- c
a>u
>Ci
• *t*-* (/>TJT Ccc«C 0
^ fci CU^i-l-
Ou«-
— O o*.>— • TO
^fr
"U f ft;C t *-(1 *- w"
*_ O X•*- Xt/»tf y»-^ ra
0£ u~j •- a-LJ >a<- c*-- ctea..*~i •
C. STr. « c
a, t- u.at«-^i L.2 V C oT" • .•* OX1 U OU
u
•»if" •»c. —*- -ic c.i- i:
*- >.Vr
f—•
— > ••G
-C
P-«.i,i *
t X-3_
if. C,
^• k.
T: wc. t
L,»
0 »
r- Cr
•.r-
V r-
•:*^--
~ iU 1 ( -
>
U
c
tfl ;TI u C *C Q ft'** .o (••> C.T: •
Se c ?"• *~ •L. C) (T *O X *
t.*. ^^1 *c. t a-i Ji TT T «U C S-u. r T L.3 t. GV -c. *-* * "~l • • ' " - " • «
O *-*-- 0, » *• *•- 0, I- 6. •L. > i. " *
t — ••- J- r i- -> ^ t< »* c *- *
.»- k. u -P- *i,-*- C WT •- _ ••
U — I- - t' •
41 cr T.'*- _: -r: o n 3 c ^ *
t. — • ., «» C <-•-£. '• -L. o, *- C — ••: -t t" • - ' * - «
"C'^ L. ^
C *• C. 0 C r. .c. • • — •t- t c— c '^ i> c •- »" :* •i »' O ••
- *" "^ '- *r- 3 I U '_ -
i_ u L -'"- »- •i-< C -^
»- T ^ C " C •ire •*- ^ * *C T "*•l- L- •
U ~ •- 'J •
•— L . ' • • - >
- - ". •
-
•p- »
,. •
0
c
i ' i i i i i -f i *• ! t r • i i t e- i i i c* i i i l i i i I i i t t i I i ) i 1 I i i i i t » i i 1 i i i I «- l c i i i i i i i i i i i i i i i i i t i i i i' ' "-" ' ' f i > i i I i i t t t i i t t i i i i t i t i i t I i i t t "• l l i t t i i i i t i ( «— i i i t i l i i i i i i i i i i i i i i i i i i i i * i -•
• - .; ,. . — .- - --, - - ("- f' . —^ - , -r• • • - - : ,- -rv—"T.'Tr* <~ •'- TT C^ -" '" -- .-•—' r.'f C'C"t~ T' .*/( vr.-C/C''" CT"."' Tf 'C ' '"-C T "** -"C '" ." ' ~. •" •' ~
•• - : ; ,' ; ^ ; : r-c-cr r c- r c c : : c c c c . c"f cTcrr c r :-c . c r..r w c c.<r r .—r-c~c rr c r.r c~c c t~c j r-r-c. c r^ r c r c , r . . ; -r ~: :
V fc J, ^ t t • "<*
112
bOr
U
•cc
c
c4,a.2
C
i
£
*"
•f
(-
. ,
>— '' '-
^
.-
. .
.-
r r
i* !;i;
„-.
k kU 1 *»- A>- ~ e• •& >J 3
i. a< A* C* O Jt
... « 0. 0», i- -. ~ C ».u « " o r4, e c ••• oX O O » — v>
II U. < * « , ! . Ul
VI t ^ tt&. i^ (Q O li I/*— o a j"^
a. k. vi 3i. i> ^» •a> > » »- o3 — T: co &. • * "^ *^ t
c c: u c- it ><u I r ' i r.' > aj£ > C "3
** < j( C ,u>o n & c c 3 •*"~ O '. 3 •- «•
4 L. 4~ I. l-» •.. ii Ul i. » f •~ 3 -^ » C * 4' C
U C C . — 0 «• t- •- C
-'" a^ "~ teA* t»— V^u •— ^. T* X •• *••
• c *: ^ c ?'-' Cr
T ' T * - W CC'*-1^- f o • » OCC Wf .C <- *«* C H^ k . C * ^ ^ 1 - * - ' ^ * ^ t
•^ C 9 0.. • J« •• —i- > n r c of >-• c c---«; *•• * • " 'O _ ^ O < * J > J C . ^ -^
t. r i. r» r cu c t'c <- c -••- n PJ*J •— fci• • > c n ~ t- u t •• >- c-'" r" j*~i "— >- o «-fc-U.**-.^ ( ' — « > — 1> ^" **C, Cj 1- t. U •• t f
r c- "»• r. f
i •*• r r. •«• >- T c ».r —O f**'
0*
r-O
t >•f-
^•c
c rr
T7r-Ot I-
Or. e3IT C
^ ^blTC. >• O
Ul -^ •*- t-e u^- fx•-••-n n
T •"t- C
O
lU
T- >••"
^N X)oc ~*"p^J»*oi. T.-
C i.-3
^ ^<Liri-. >• a/
I- c~^rf •*~ u
(• uh- C.^-I-
1- 1:
^ "t- t
l_
QC
C
**
-
O Cu•c^ to ••-
0 >.EC s.3 ••-i* E fc1*> 3 Ca.«' oO 9 w
• S. 1-^ ^ • ^4- . O I:
••• » -r- -T "t-: u C— """ CK-C; O r '->r
II II I- .-; t-
V'.'tf V (~ *h r- o
a/
f> Ck. E
"3 »C• i
r .* rK- u -I; *
O •-
H U --t E c1/1 a " *•
— tn t-(_ x t-
(/> E ilI- ..- i.•* , b ^
~ •- X
»- TT C' —trt Cl w. r:••• r-. «- »-i. >. i' »
C u. 1.O i . 3£ r™i •» O•• » — a. t-
^ X '^ <"^** t ft* ^ ->
0 i» b u« ii o x^ e i< —« TI ».«-^ e 9
f l ) ^ - n f c * X H - t »• • a < - i i o t . x «~ l_l C •- X— .. <0 (-1
> •• C o j. C — c . 0> i—
& r 3 r.c-.» ,1,(L ^ O t> *^ 'O X *• *-
* kl ^^ Oi/1 tr ^* C O — • •« Q
». O II "(-I •- 0; .-- IS C TT
c </> a * ~ u a / c a/ ra^- c •>-»-»- c o ; 1
ui -^ f r^ C>«- 3 ~c —a ^»>
<B c. c/i k o» e.*- •• <^t.' •- ^ n — *- 1*- -. t! u OCO »i£ II T3 •• —•••->-
K ci- ** ••- "5»-iii (= •..«.. o» *-:-i-.
•^ '&O « t O « *rf»'^' M - f - C *rf»rf*'»' U.k .
C w t-*Cr— *^ c. Ctt (->— » .^-.^•»- c.. V!_•• - •»- i_ —• c«- .T«—^; - w c ' : • - • • »- '" ••
9*^ E x f t i f l ' T'OC^1 XlCi » . ^ _ « ^ •*" t.c t- n i - c . i i < c c . — cat to* L: o c-c •• .£.•»- cj; c *-•- r;-^
u e •->- -a o —— 01- T. C >- 3— & ci <- n <** °«:' 4^ « • «-»' CO/ II £ *-C «-T; ..-,...-
1. T T1 —i I- •Jr. c. r c
r r i-».- * ». .-
i » i i • t i t i i i i i i t i i i i i i i i i c' i i i i i i i i i i i i i i i i i i i t i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i -'••< i i i i • » i i i i t t i i t i i i i i i • i i t :.•> i i i t i t i i i i i — i . j i i i i i i i i t i i i i i i i i i i i i i • • i •• i t i i i i i i i • i
, r ; r ; c - r r^r r c~c.c r rc~cc r-c : : r c c r-r C~T r r •- r*c : r-c r - : c .-•; : c c ,. t -f c-r rr : r r rrc^c c-.r r c CT c.c r".^r ». r c r
C 1 • t/ ^ t C • %^ 4^ • •; t- t/ »., » « , * , « _ , *- • '•
• • * w *
113
o
•» c. »-• o— ' »•- i* cr
"- -J t. vJ •*
» c, L_ • i_, *-•*• • "~ •» • •+•
* »• c, r ' t-. C,r*- L- ••' •— C«* 0
~i: I L '; 7H z
r "^** !'.'; ^ ./x* "~ u
• - V * . C .IT J X X C C .. V-f• — •• t c t- »•»*« e. a. .r *- -^ G*
•»- ••- r — cJ K c — £
k
***
'(A kO >
— uC-
ko — •-
l-« 3 E
>- t *"-*k i C' U> >ftl 1 •* k
w W (0*«k l/l »> c
O C «-T-«- X
• «v » » 0
T^ C X4«*k. o -^ V* c c *^
> 1.1 in *> c C •«— •or, o o a. cc• / «- k II (~
17 X •• • — Ci VI
h- — ' — XM
o t- — c c ar. k ct -* k> x.•3 «--^-«- I3C.W k •«
.;— o- o o c»- i- i/c
*J ^"••™
0o•Dax
• *«»ri Xr: «
r— »•I_J Q/i-0 •«ft*t Xw Oi t)
C.x —^'^ •» Or
« <o c **
=! k k
e «j t •«k cio:«- a*c c E
~" «v— * oII X.«k>
s. oX ••<C II II
0. C.C *•"•-
k ^-_ .4. ZOOe i: i-i'
f
r~ i
kj
^ ~
^ •
(C ^**^»•-*-•
fc-t- «fl
C < 0.
li <O <C.. ^- u V.
X*-"-> Q. <->
— —-J L. C
^^C o**' c
f. C •« •"- -»k> f-C C.7: k r- — f
C C 0 i^ ^
C "Cc •»C T-
C
T3Ct
•rejrkOJ
a>c
tfl.^-4 * *-
tT *•«•. f"0
O W^
CJ >
C ', C.^.— cr. r *'
C' *-f-^4. r3 t» a
i-c r
l/l Xu -*tf> ^ (Q
•» tf\ ^•*• \ c.\ "Om •> x-«-
£ £ K
•S - *. - i-1• w> c C
c- a. 0• X k ' ••-
Otr k inrro 3 •>»-3 u» ••"
1- E« k. • m
;- > cb >eCK •• *-
••-rj «. »-<-'c; e«^ *v» ^ in
-. c>^ ~ 3C*- i. «- i«-^3 l-C C W* *"C 3^' ^ 6.''-
< «.« 330
c a *e -~
•^ ur c c. c
ID T- h-
~ r^.,!'. ~ ." -^ - -
C f » - C . 3 3 3 = 3 => «- >C S.--C C.- C.
^i- .- » 3 3 3 3 -5 1T-
~C 0 ^"_l»^ C C C C C CC»- ir- C-— -•—a e " r H - i : c t a . t c c c
k C 3 k w k ••-.•••• •f--^"»-"-
kc ••• t r» _ic :-r i-r r :.> j.
• » »•^» ul •*
Xtiu
— — o«< "i n- o o3--1-
t p
in t
•;•S
SJ
T>
f lO
St
^ O C
s .
• %
f r f rr -. ir
r c c -"c- c- . -
: : -
>
ct
t.
,
''
'
..
-
1 l| I I I ' [ J I I I >••• I I I r • I I I I I I I I I I I I I .? I I I I I t I I ••' r i I I I I f .«- I f. I. I I I I I I I I I I I I I I I I I I I I I I I I I I I I I1 '•' I i ' I -i I I I I I I I i I .-i I I <•-. I „• I I | | | i i i i I I I | | I | I I i j I | | I I I I I | | I t I i | i | i o I I I I I I I I I ' I I • ! 1 '
< •" -j -: r i\r o
• r-.; r re - re-— s-i" c. re J'r — c -r r-e r c re-fc-c-r rrrc c r ~r-r. :' rr s-r i- r r :-r-c c cc- c r rr rc
•rr r-p c~« r -K vv »'r e c : •• r t- r -\ f"r-1 ••
t t c t
114
«i T3
C x «a >- r-
a, "-3
% * *
0,
c, -^> c
u "
•*. L.
— ;. o
C •" T
» » \
;• ," ."
, — — .»=; .= 3 = =» x » •» >
,* - 1 "3 ^ 3 *-
-• c ti r c
; ^. .- ' t ••
: : i r :. .,
\ .. t>0 J,.,.. 9
\ c xt~ a.
» X"---«- O —
«— X O *B O O
-» c c> O O 4« t. •»- >« C Ifl ••-« 3 (A *- *-
u: % C o~ i- i- r
C C *• f T C7* *^E to
C. ••- — Wl C w1 — 'r »- 3 c ^ m <8
^. »•» ™ c o•• Oi c,.^«-< —
~v r : M'-f.-Zu'licauiuS
=r^=.'^|>=^,Vc c c r c c c c c c c c c
.c;.tiii: ::::t '.c;.t.-.r»S«3«=.-« =
Pe
rCc
ntA
ct
-
••
E
0
f
r
i ZJ
3cr
C
ir
TJ
K *%
VI
••• 3
t Xf
•c
ti
u
o
& Q;
X *-i.- O
•+-
^ ^
O 4-»" f
**•*
*/»
a
*•* .* p
3 3 3 3 *..C. ;i C. C. *I
•0333 >-:0000 i;
c c c c c
a. c c, c
CC'
o
or?
>« -.* Of
3—*t'«-
o a
•vU O
•3 12 t
C3 a
3 -*C".-C'^xa> x
•*-«^ n>
— *- o
5«2i-«- 3
X i. :.•
o
.rf •• % r*•^ . \ \f ••O » 1
u \ %o. • ..
• « t k »
u ^ O-"CL « ZO
- 2 .: £ :;>'_3 OC C3* >™UiO O
L. }" •••r j trt\ t i*-t- 3 \ > Z 1^ •
•^ i- * * O *T •% C
3J ,- t- M « ..« % u
* »-% wo*-*z oc x: •
£. . C* C \ 1 • V- t W •K •* uji: ft *t~ \ ^ . » o u w >>-»- oo » r •- or,^' T i. uD 10 ^-(j ?• rr »^*r- o r1-^ >
» - - c r « - - t - c * c * t o r e
JT «"«" ". 7".ls ' .^t ^°:rn •.
C » » r > f K > « * « c x r 3 3 r * - K 3 3 3 * « -JC ' - *^ - • -> 1 T 3 ^ 3 3 3O Cv.jCJ -< 3 OC "J* 3 3 ~ r- i; '-. C r1
ur — t c r c c c c c * «"* • • - »"c «.*••• c c ..^t.—
> £.
•* i^ c
4->-lw
C »
c *-
I I I I I I I t I I I I I I I I I I I t I I t I I I I I I C-I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I: i ' i i i i l i i i l I l i l i i i i i i l i i i i i l i i l i i i i i t l l l I l i o l l l i i i i I l l I l l I l ' i i i l i I i l I i i i i i • • : i
r .» r c r- r-r> ^ t-
w c - .- r r r r CT r ." r-c c ~c rrr- rc: c c~r c.- "r-c c ..-f" t" cc. ;• c : r r r c-c.rc?c~c crcc C c-c c~c : c c r : :
r f -.jtr.-.- r_r ^r-f'r~*ru~^:r- r r :'*• nr «TV-VC- fC frr.r «•«••<•£_£? <~«-f .»-ytn.cf r c r r-r»^*ttr'Ct c-c_Cr-ff'f<.—~<' c r r r r - f - f u - - ' f :
• • • • I t ' t- IT t t, t •
» • »
115
o o *— fliO O w 31w- «- C. w
C.
t> O O *^ "*e o •*•» o > >** «- OCJ *- fc. t~p- • • C *•• C **— « « OO * *• C* C-* *«OO C £ «-•£ »-»
oo £2% Z* % O •»• *-O **J «4 V»» 3 T> U«-Oifc^ C
* * % = ; • oc% % a* < > >o «-
4* >*** icc *:» _j • • (t a: o a* e* « *- K t, »» c c n . . ^j . .••- •— *-» % •-. •• *.j % h- - o < ^ - * £ e Q * * * H - « — a-*-1 a» » ^ j * - ^u *— \ i^«»>- » t * o w^^ *O • *- •*» 6"%^» C ^ E * 6 u J W « - wr 4.** c. \ * no % * % r* c -i- •»• *- *- »•- *»*^- •»• T: ••<".E *«>- 3 ;-« ^ * ^\ irt-7 £O 0. >- W * * *~ *-** h- ^ .. .. u^c i r c » * » - . « •»» a , k ~ * * - * > e o a* ^ — > - - i > » —
*-\ 3 LJUJ fc'^ -*— » * -CT- O U 6- *-O .^^k-r-'^** C k. f U .^ I t/ -^ C*r <-?o » »u; » *-<: ^» "c c c- ir * - f - > - **-c me o * * « * ^ - * c - — < • • •
ZT »-i >% CjE U.I-H . *• •• c 3tr" >- A w 0*«-— U >- C UO »-* CC" o <• K O *J*-u. *- >u- u. LJ •-• >4_"»^r'* * "C rn O> <£••*—• O >**^ ^* •* • • •»«. ^- ^>CTt. C L..O. Cf »«H- t-. -.% * 3 •» "i crt C M QJ >••- TS O iT- H-U U II w~ II II C~-OC".C- I •• w w •
........ • *> •-. c c -i' «" " cj i* « c~uj ^#•• •* > > > > .... c » « -r
«-»C) o. •- c c M u »- * -* - r ' r
O *-• v-O «-»>-^( jx . 'OC^ '«JO»J^ 'UC»O* '*JC£«j-^j c t;cw<tC7i>-«C' c tit. c. ttt- t. fc^
— •- — — *-~J x *^^^* • »*-.* %\ » % \ % %\ » « % r c a . t»«» c > A t. c_c. C"- c •" c •*• c
C C" C f * . «. ^* *^c *- *-> *-c ^- 4^ *-O +* * K x x w 2 3 3 f c 3 3 r « 3 3 3 x *CO»-" 1 " 3 " * ^ 3 r c ^ : C 3 O C C 3 O O O 3 — * - i :<CO*fT«f«^.*^^)r«w«. X X f c . t « K X « X U •• <*-
.*^^v^^-v v 3 3 3v 3 3 3^ 3 3 3^ «( » ^T-^
. k k L. t. t. k k k L . k . L . U k u C
'' I I j I '- ' I I I I ' I I I I I I I I I I I I I I I I I I I I I t I I I I I I I I i I I I I I I I I I I I I I I I r, I I I I I I n I t I I I I I I I I I •< < < < < < I < I I <<> I I > I I t I I I I I I I t I I I I I I I I I I I I r- I I fi I I I I I I I I. I I I I I I I I I I r'l I I I I I I I I I I I I I I I I I
crc c : ~ * - c c ~ c c r r c r c c ^ r c T c r cc-c c r^f r .• r rc - ' : rrrr .~c.c-c~c c~r c cc'cc:» crcrr-c c r c~c cxc~-c t c c^c-c e r r : :
»t ir v: r y r C"»-r r ^TI. vr :" r *,""*• r >K^OL--<''P c*c C" r-r .p'U U--C-P CT r'9-r-.f^.-^fu- vr c c f«- r ^" *J i" -' * » .'.- v. ,;»^- _-...- ,-,j- .-*.- ^- j *~ s *?.•*•* r ^ * r-r- fi-p r^r-, .r -- .••-*•?** *•.- j:c r r c r r r - c* r r .* -~. * '" r ~ •.
^ - t ~ i - t - i < 4 * * J
116
51
«i.
a:
o
o
c*
•'•",_,
o>c
cin O "C -~ asc .» ~ c
••• X O•* ^- k tl
~ «l -0•v x a a/c • c o o .o r uU • «/ X
k 4-» O»/*- *- •» •* •* o
1- .* C «- •< Clc T u o — wek I/I J .- -^ |- X .->C .' C. 1. TC T Oi "• *- II ra — •L. ~^ J* k 4>J II ^ •<•
r. o a* u •*- +4 +*u > % f ^ t A £ o t n c
c c/ -^ * >- e r«- ^ f- u. r- »•- — ji
c: c — *• a> x ^ oi f o u r - i c j » -O * ^ k C T 7 k . O J C i
cr— a «• c «. r c.— C j: t •" 3 <bC~- w »- c.' f u. c* W Uc
-* r- 1
£ Cre C£ U0 0L> ^"a <~
cS oc• ^ 4 , «•
• * (ft i •*-^% «^ ** *C;
u — C*- — c ow O 0 U
e <n «K C "• -»'
— ' W- a. — cu *- o —> U U «.Ju -^ < a<r 3 c -
«.. ^ c o t•* ia c o< o.u u ^i w
p ^^ *^
J,"^_
Xft>;;•
U
*>p>
•» C
i.w ii._
e c>~ c•* ^~K M0 C~ tl
- c••• 3k V-
C
T «-•C o/ v
*iA'C
IIA
X •»-• X
i- c~ o
a> oC w
•co
c >^o
t- E3 C
C X r C
t. re r c>—r c v\*"k e ~i-.O 3
C I.— D
C »
c c
To
O k U
* r -<•
i i ' i i i i i i i i i i i i i i i i i t i i i i i i i i i i i i i i i i i i i i i i t i r ' i i i i i i i i i r i ' i i i i i i i i l i r' ' I ' I i I ' I I I I I I ' I I I I I I I I I I I I I I I T- I I I I I I .-I I I I I I | I | I ., I I I I I I I I I I I I I I I I I I I I I
r t~**f'f- '*t*f^*fiif*j*£*f± ta *AL '^<'*'**v'^^^-^ ^-c ^ .c_^.r - r -»-^-»--»-r i r>-r"K^f>*/vr*/ t^w
•-f-».^ *^-f *-*-•' *'.<*/ ^f «-*N/»f x.*»f */NTvJvf
r r r ^ c r r c r - c r r c r r - r t c-c. ec-ccrTrT'c-Pc-F-rV re T-- c"F.' r r r r r r c : r c i c c c c r rrr r cr
oc
i
i,*>o
r.r:o
3 ~^ f
aco
. 01 . . -•. c »-: