computing types

8/19/2019 Computing Types

1/20


2/20

Distributed computing:"istributed computing is a computing concept that, in its most general sense, refers to multiple

computer systems wor-ing on a single problem. !n distributed computing, a single problem is

divided into many parts, and each part is solved by different computers. #s long as the computers

are networ-ed, they can communicate with each other to solve the problem. !f done properly, thecomputers perform li-e a single entity.

The ultimate goal of distributed computing is to maximize performance by connecting users and

!T resources in a cost)effective, transparent and reliable manner. !t also ensures fault tolerance

and enables resource accessibility in the event that one of the components fails.

Advantages

$hare ability

%xpandability

ocal autonomy

!mproved performance

!mproved reliability and

availability

otential cost reductions

Disadvantages

0etwor- reliance

Complexities

$ecurity

+ultiple point of failure


3/20

Parallel computing:Parallel computing is the use of two or more processors (cores, computers) in

combination to solve a single problem.

1. Types of Parallelism

There are various forms of parallel processing1

) Bit Level Parallelism, each bit of a word is processed at the same time) Instruction Level Parallelism, execute multiple instructions simultaneously

) Data Parallelism, (a.-.a. loop level parallelism focuses on distributing the data across

different parallel processing units

) Task Parallelism, (a.-.a. functional parallelism focuses on distributing execution

tas-(code 2 data across different parallel processing units

2. Types of Parallel Computer

3sing &lynn's Taxonomy, computer architecture can be divided into1

) ingle Instruction! ingle Data stream ($!$"

) ingle Instruction! "ultiple Data streams ($!+"

) "ultiple Instruction! ingle Data stream (+!$"

) "ultiple Instruction! "ultiple Data streams (+!+"

Today's parallel computer are all +!+" type, in more coarse)grained style, parallel computer


4/20

can be further divided into1

) P"D $ingle rogram, +ultiple "ata

) "P"D +ultiple rogram, +ultiple "ata

#ccording to the memory architecture, parallel computer can be divided as1

) #ared "emory Computer

!n this -ind of computer, each processing node share the same global memory address space.

rogramming on these computers can be as easy as on multicore wor-station.

$hared memory computer is easy to programming, but since bus is used among all

processing node, the scale is limited (usually several tens, as bus contention will become the

bottle nec- when scale arises.

$hared memory computer can be further divided into two -inds1

) 3+#/cc)3+#, all processing node share the same physical memory device through a bus

) 03+#/cc)03+#, each process node has local physical memory but accessible by other

nodes, but the access time depends on the memory location relative to a node

3niformed +emory #ccess (from 456 0on)3niformed +emory #ccess (from 456

) Distri$uted "emory Computer


5/20

!n this -ind of computer, each node has its own private memory address space and can't access

other node's memory directly. 3sually, processing nodes are connected using some -ind of

interconnection networ-.

"istributed memory computer can scale to very large since no bus contention occurs. 7ut it's

more complicated to write program on this -ind of computers.

"istributed +emory (from456

) Distri$uted #ared "emory Computer

!n this -ind of computer, their hardware architecture is usually the same as "istributed +emorysystem, but its interface for application developers is the same as $hared +emory system.

"$+ is usually implemented using software extension to 8$, which some performance penalty.

Advantage:

• $ave time and/or money

• $olve larger / more complex problems

• rovide concurrency

• Ta-e advantage of non)local resources

• +a-e better use of underlying parallel hardware

Disadvantage:

• +ore hardware reuired, also more power reuirements.

• 0ot good for low power and mobile devices.• 8nly certain types of data are suitable for parallel processing. "ata that relies on

the result of a previous operation cannot be made parallel. &or parallel processing,

each data set must be independent of each other.

Cluster computing


6/20

A cluster is a type of parallel or distributed processing system, which consists of a

collection of interconnected stand-alone computers cooperatively working together

as a single, integrated computing resource.

In cluster computing each node within a cluster is an independent system,

with its own operating system, private memory, and, in some cases, its ownle system. !ecause the processors on one node cannot directly access the

memory on the other nodes, programs or software run on clusters usually

employ a procedure called "message passing" to get data and e#ecution code

from one node to another.

Types of Cluster

$. High Availability or Failover Clusters

%. Load alancing Cluster

&. Parallel!Distributed Processing Clusters

AD%A&TA'()*

5 !t helps in reducing cost.

9 !t improves networ- technology.

: !t is always available in all possible

situations.

DIAD%A&TA'()*

5 !t is difficult for developing software for

distributed system.

9 !t can be easily accessed and applied to

secret data.


7/20

"rid computing

;rid Computing is a type of parallel and distributed system that enables the sharing, selection,

and aggregation of geographically distributed


8/20

"reen computing

• ;reen computing is the practice of using computing resources efficiently.

• "esigning, manufacturing and disposing Computer, servers with no impact on the

environment.

• To reduce the use of =azardous materials, maximize energy efficiency during product>s

lifetime.

"anufacturing of PC0s

Lead) used in soldering of printed circuit boards lead can cause damage to the central

and peripheral nervous system , blood systems and -idneys.

"ercury) used in batteries, switches. +ercury spreads out in water transforming into

methylated mercury that can cause chronic brain damage.

Cadmium) used in resistors for chips and in semiconductors. Cadmium is classified as

toxic, these compounds accumulate in the human body, particularly the -idneys.

'reen manufacturing

Bam$oo) is becoming increasingly popular for ma-ing casings for computers and

peripherals.

,ecycla$le plastics) computers are constructed from non)recyclable plastics i.e.recyclable polycarbonate resin.

(co*friendly flame retardant) there are flame retardant silicone compounds available

that are flame retardant and completely non)toxic.

Inventory management) reducing the uantity of both hazardous materials used in the

process and the amount of excess raw materials.

%olume reduction) removes hazardous portion of waste from nonhazardous portion.

(nergy use of PC

C3 uses 59? @atts

CT uses 5A? @atts

B hours of usage, A days a wee- AD9 E@atts


9/20

,educing (nergy Consumption

Turn off the computer when not in use, even if Fust for an hour

Turn off the monitor when not in use

3se power saver mode

3se hardware/software with the %nergy $tar label

"on>t print unless necessary and you are ready

3se C"s instead of CTs as they are more power efficient

ave energy #ile orking on t#e computer)

leep mode)

$leep or standby mode conserves energy by cutting off power to your display, hard drives

and peripherals.

#fter a pre)set period of inactivity, your computer switches to a low power state.

i$ernate mode)

• $aves energy and protects your wor- by copying system data to a reserved area on your

hard drive and then completely turning off your computer. educes wear and tear on your components.

'reen Disposal

,euse) donate your computer components

to people who may not have or have lesser

uality computers.

,efur$is#) rather than discarding your

computer upgrade it. Change its some of

the parts in order to ma-e it new.


10/20

,ecycle) 8ne of the maFor challenges is recycling the printed circuit boards from the

electronic wastes. The circuit boards contain such precious metals as gold, silver,

platinum, etc. and such base metals as copper, iron, aluminum.

$i3uitous computing

$i3uitous computing (u$icomp is a post)des-top model of human)computer

interaction in which information processing has been thoroughly integrated into everyday

obFects and activities. The idea that almost any device, from clothing to tools to appliances to cars to homes to

the human body to your coffee mug, can be imbedded with chips to connect the device to

an infinite networ- of other devices. efers to the use of computers in everyday life,

also called pervasive computing

'oal of is u$icomp1 to create an environment where the connectivity of devices is

embedded in such a way that the connectivity is unobtrusive and always available.

Tec#nology %ie

Computers everywhere G embedded into fridges, washing machines, door loc-s, cars,

furniture.

!ntelligent environment.

+obile portable computing devices

@ireless communication G seamless mobile/fixed.


11/20

ser %ie

!nvisible G implicit interaction with your environment.

#ugmenting human abilities in context of tas-s

C#allenges of u$icomp

1. T#e 4Accidentally5 mart (nvironment) !f you wal- into an environment

anywhere in the world, you would probably not find an infrastructure suitable for ubiuitous

computing devices.

2. Impromptu Interopera$ility) That means many technology)producing companies

desire to produce their own proprietary products spea-ing their own proprietary language.This leads to HnoninteroperabilityI between devices from different companies.

6. &o ystems Administrator) "epend on the challenge two there is lac- of a systemsadministrator.

7. ocial Implications of aare tec#nologies) 3biuitous computing will have a social

impact on our society Fust as the previous two eras of computing did. &or example1 =ow

will an individual -now if they are within a HsmartI environment where embedded

devices are gathering dataJ !s it ethical to gather information from individuals without

their -nowledgeJ

C 8 ardare Demands

• Technology reuired for 3C 1

Cheap, highly advanced $! technology (nanotech,

ery low)power computers with convenient displays,

ow)power, ultra)fast networ- for interconnection1

wireless end)points

cellular topology

wide)bandwidth range

$oftware systems for 3C applications and support.


12/20

Biological computing

7iological computation as-s how and why living systems can be viewed as

fundamentally computational in nature.

Advantages

erforms millions of operations at same time

;ood for parallel computing

#bility to use large amounts of wor-ing memory

5 gram of "0# can hold 5 x 5?5K

+7 of data

8r 5KA trillion C"s

5 C" is B?? +7

Cheaper

ightweight

5 lb of "0# has more computing power than all computers ever made

ow power used to -eep in original state

=as ability to solve hardest problems in a matter of wee-s

%nvironmentally friendly

Clean, readily available materials

Disadvantages

• +olecular operations are not perfect

• "0# computing involves a relatively large amount of error

•

#s size of problem grows, probability of receiving incorrect answer eventually becomesgreater than probability of receiving correct answer.

• $ometimes there are errors in the pairing of "0# strands

• $imple problems solved faster on electronic computers

• =uman assistance is reuired

• Time consuming lab procedures

• 0o universal method of data representation


13/20

• "0# has a half)life

• $olutions could dissolve away before end result is found

• !nformation can be untransmittable

• Current "0# algorithms compute successfully w/o passing any information from

one processor to the next in a multiprocessor connection bus.

Cloud computing

Cloud computing is distributed computing on internet or delivery of computing service

over the internet.

Cloud Components

!t has three components

5. Client computers

9. "istributed $ervers

:. "atacenters

Types of cloud

ublic cloud

rivate cloud

Community cloud

=ybrid cloud

Cloud ervices can $e dived into 6 stacks)

aa 9oftare as a service:1 euired software, 8perating system L networ- is

provided.

Paa 9Platform as service:1 8perating system and networ- is provided.

Iaa 9Infrastructure as a service:1 Fust 0etwor- is provided.


14/20

Benefits of cloud

educe spending on technology

;lobalize your wor- force on the

cheap

educe capital cost

!mprove accessibility

!mprove flexibility

ess personal training is needed

+onitor proFect more effectively

#chieve economic of scale

Disadvantages of Cloud Computing

• euires a constant !nternet

connection

• "oes not wor- well with low)speed

connections

• &eatures might be limited

• Can be slow

• $tored data might not be secure

• $tored data can be lost


15/20

Different security related issues of cloud computing

M +ost security problems stem from1

G oss of control

G ac- of trust (mechanisms

G +ulti)tenancy

M These problems exist mainly in :rd party management models

G $elf)managed clouds still have security issues, but not related to above

•

Loss of Control in t#e Cloud

M Consumer>s loss of control

G "ata, applications, resources are located with provider

G 3ser identity management is handled by the cloud

G 3ser access control rules, security policies and enforcement are managed by the

cloud provider

G Consumer relies on provider to ensure

M "ata security and privacy

M esource availability

M +onitoring and repairing of services/resources

•

Lack of Trust in t#e Cloud

G Trusting a third party reuires ta-ing ris-s

"inimi;e Lack of Trust) Policy Language

M Consumers have specific security needs but don>t have a say)so in how they are handled

G @hat the hec- is the provider doing for meJ


16/20

G Currently consumers cannot dictate their reuirements to the provider ($#s are

one)sided

M $tandard language to convey one>s policies and expectations

G #greed upon and upheld by both parties

G $tandard language for representing $#s

G Can be used in a intra)cloud environment to realize overarching security posture

M Create policy language with the following characteristics1

G +achine)understandable (or at least processable,

G %asy to combine/merge and compare

G %xamples of policy statements are, Hreuires isolation between +sI, Hreuiresgeographical isolation between +sI, Hreuires physical separation between

other communities/tenants that are in the same industry,I etc.

G 0eed a validation tool to chec- that the policy created in the standard language

correctly reflects the policy creator>s intentions (i.e. that the policy language is

semantically euivalent to the user>s intentions.

"inimi;e Lack of Trust) Certification

M Certification

G $ome form of reputable, independent, comparable assessment and description of

security features and assurance

G $arbanes)8xley, "!#C#, "!$TC#, etc (are they sufficient for a cloud

environmentJ

M is- assessment

G erformed by certified third parties

G rovides consumers with additional assurance

"inimi;e Loss of Control in t#e Cloud

M +onitoring

M 3tilizing different clouds

M #ccess control management


17/20

"inimi;e Loss of Control) "onitoring

M Cloud consumer needs situational awareness for critical applications

G @hen underlying components fail, what is the effect of the failure to the mission

logic

G @hat recovery measures can be ta-en (by provider and consumer

M euires an application)specific run)time monitoring and management tool for the

consumer

G The cloud consumer and cloud provider have different views of the system

G %nable both the provider and tenants to monitor the the components in the cloud

that are under their control

G rovide mechanisms that enable the provider to act on attac-s he can handle.

M infrastructure remapping (create new or move existing fault domains

M shutting down offending components or targets (and assisting tenants with

porting if necessary

M epairs

G rovide mechanisms that enable the consumer to act on attac-s that he can handle

(application)level monitoring.

M #d#C (is-)adaptable #ccess Control

M + porting with remote attestation of target physical host

M rovide ability to move the user>s application to another cloud

"inimi;e Loss of Control) tili;e Different Clouds

M The concept of N"on>t put all your eggs in one bas-et>

G Consumer may use services from different clouds through an intra)cloud or multi)

cloud architecture

G ropose a multi)cloud or intra)cloud architecture in which consumers

M $pread the ris-

M !ncrease redundancy (per)tas- or per)application


18/20

M !ncrease chance of mission completion for critical applications

G ossible issues to consider1

M olicy incompatibility (combined, what is the overarching policyJ

M "ata dependency between clouds

M "iffering data semantics across clouds

M Enowing when to utilize the redundancy feature (monitoring technology

M !s it worth it to spread your sensitive data across multiple cloudsJ

G edundancy could increase ris- of exposure

"inimi;e Loss of Control) Access Control

M +any possible layers of access control

G %.g. access to the cloud, access to servers, access to services, access to databases

(direct and ueries via web services, access to +s, and access to obFects within

a +

G "epending on the deployment model used, some of these will be controlled by the

provider and others by the consumer

M egardless of deployment model, provider needs to manage the user authentication and

access control procedures (to the cloud

G &ederated !dentity +anagement1 access control management burden still lies with

the provider

G euires user to place a large amount of trust on the provider in terms of security,

management, and maintenance of access control policies. This can be burdensome

when numerous users from different organizations with different access control

policies, are involved

M Consumer)managed access control

G Consumer retains decision)ma-ing process to retain some control, reuiring less

trust of the provider (i.e. " is in consumer>s domain

G euires the client and provider to have a pre)existing trust relationship, as well

as a pre)negotiated standard way of describing resources, users, and access

decisions between the cloud provider and consumer. !t also needs to be able to

guarantee that the provider will uphold the consumer)side>s access decisions.


19/20

G $hould be at least as secure as the traditional access control model.

G &aceboo- and ;oogle #pps do this to some degree, but not enough control

G #pplicability to privacy of patient health records

•

"inimi;e "ulti*tenancy in t#e Cloud

M Can>t really force the provider to accept less tenants

G Can try to increase isolation between tenants

M $trong isolation techniues (C to some degree

G C.f. + $ide channel attac-s (T. istenpart et al.

M Oo$ reuirements need to be met

M olicy specification

G Can try to increase trust in the tenants

M @ho>s the insider, where>s the security boundaryJ @ho can ! trustJ

M 3se $#s to enforce trusted behavior

• 7ig "ata1 # definition

• 7ig data is a collection of data sets so large and complex that it becomes difficult to

process using on)hand database management tools.

• "ata come from many uarters.

– $ocial media sites

– $ensors

– "igital photos

– 7usiness transactions

– ocation)based data

•

• The four dimensions of 7ig "ata

• olume1 arge volumes of data

• elocity1 Ouic-ly moving data

• ariety1 structured, unstructured, images, etc.

• eracity1 Trust and integrity is a challenge and a must and is important for big data Fust as

for traditional relational "7s


20/20

•

• 0o$O (0ot 8nly $O1 "atabases that Hmove beyondI relational data models (i.e., no

tables, limited or no use of $O

– &ocus on retrieval of data and appending new data (not necessarily tables

–

&ocus on -ey)value data stores that can be used to locate data obFects – &ocus on supporting storage of large uantities of unstructured data

– $O is not used for storage or retrieval of data

– 0o #C!" (atomicity, consistency, isolation, durability

•

• 0o$O focuses on a schema)less architecture (i.e., the data structure is not

predefined

• !n contrast, traditional relation "7s reuire the schema to be defined before thedatabase is built and populated.

– "ata are structured

–

imited in scope• "esigned around #C!" principles.

• @hat does 0o$O 0ot rovideJ

Poins

;roup by

7ut 03T$ provides interesting materialized view approach to Foins/aggregation.

#C!" transactions

$O

!ntegration with applications that are based on $O

•

computing types

Documents