framework for proposed te and qos tests - project deliverables - * to demonstrate end-to-end traffic...
TRANSCRIPT
MB-NGProposals f or
End-to-end inter-domain QoShttp://icfamon.dl.ac.uk/papers/MB-NG/TE-QoS-D2.pdf http://icfamon.dl.ac.uk/papers/MB-NG/p20011213.pdf
Robin Tasker, Daresbury Laboratory13 December 2001
Daresbury LaboratoryCouncil for the Central Laboratory of the Research Council s
Framework for Proposed TE and QoS Tests- Project Deliverables -
* To demonstrate end-to-end traffic management across multiple domains using live Grid traffic (TM2)
To demonstrate end-to-end QoS and traffic management between the UK and USA (TM3)
To demonstrate end-to-end QoS and traffic management to CERN (TM4) - e-2-e services - separate administrative domains - diffserv, “other” packet marking, TE etc
* To understand the use of MPLS as a traffic engineering tool within the CORE SJDN (TM1) - understanding of MPLS - as a basis to configure traffic management for “guaranteed” bandwidth and QoS
Framework for Proposed TE and QoS Tests (1)- Broad Work Areas -
Using the defined test traffic
The proposed framework falls into two broad areas,
1. Policy-based tests,
1. With and without MPLS using
Basic packet marking techniques, i.e.
using diffserv and QBSS
across the extended SJ4/MG-NG test network;
2. Bandwidth broker developments using Expect scripts to make real-time adjustments to router configurations, i.e. policy mapped to scripts leads to router/network reconfig’s. (see work done an U Michigan in this area)
2. Cisco guaranteed b/w developments using MPLS/diffserv/rsvp to model the VPN
Basic Packet Marking Tests - Configuration -
test data
test data
test data
UCL
Manchester
Rutherford
Administrative DomainAD(C) with Policy P(C)
Administrative DomainAD(M3) with Policy P(M3)
Administrative DomainAD(M2) with Policy P(M2)
Administrative DomainAD(M1) with Policy P(M1)
SJ4 Test Network
Basic Packet Marking Tests - Test Proposals (1) -
For testing between, for example, Manchester and UCL
and for each specified Policy - P(C), P(M1), P(M2), P(M3),
without MPLS enabled in AD(C), and with MPLS enabled in AD(C)
the following configurations are required for ,
1. diffserv tests
1. diffserv operation in AD(C) only, and 2. diffserv operation in AD(C) plus AD(M1) and AD(M3)
2. QBSS tests
1. QBSS operation in AD(C) only, and2. QBSS operation in AD(C) plus AD(M1) and AD(M3)
test data
test data
test data
UCL
Manchester
Rutherford
Administrative DomainAD(C) with Policy P(C)
Administrative DomainAD(M3) with Policy P(M3)
Administrative DomainAD(M2) with Policy P(M2)
Administrative DomainAD(M1) with Policy P(M1)
SJ4 Test Network
Basic Packet Marking Tests- Test Proposals (2) -
For the test data flows where, for example, Manchester to UCL is being tested,
Man -> UCL;Man -> UCL; RAL -> UCL;Man -> UCL; RAL -> Man;Man -> UCL; RAL -> UCL; RAL -> Man;
Man->UCL; UCL -> Man;Man->UCL; UCL -> Man; RAL -> UCL;Man->UCL; UCL -> Man; RAL -> Man;Man->UCL; UCL -> Man; RAL -> UCL; RAL -> Man;
Framework for Proposed TE and QoS Tests (2)- Broad Work Areas -
Using the defined test traffic
The proposed framework falls into two broad areas,
1. Policy-based tests,
1. With and without MPLS using
Basic packet marking techniques, i.e.
using diffserv and QBSS
across the extended SJ4/MG-NG test network;
2. Bandwidth broker developments using Expect scripts to make real-time adjustments to router configurations, i.e. policy mapped to scripts leads to router/network reconfig’s. (see work done an U Michigan in this area)
2. Cisco guaranteed b/w developments using MPLS/diffserv/rsvp to model the VPN