PPT《11/14/96》

11/14/96 S.

Keshav MATE:MPLS Adaptive Traffic EngineeringAnwar Elwalid Cheng Jin Steven Low Indra WidjajaBell Labs Michigan altech Fujitsu

2006 Talk Outline MPLS Traffic EngineeringOverview of MATETheoretical ResultsSimulation Results Best of Both Worlds MPLS + IP form a middle ground that combines the best of IP and the best of virtual circuit switching technologiesATM and Frame Relay cannot easily come to the middle so IP has! Label Encapsulation MPLS C between L2 and L3MPLS Encapsulation is specified over various media types. Top labels may use existing format, lower label(s) use a new "shim" label format. Label Substitution Have a friend go to B ahead of you using one of the two routing techniques (hop-hop, source). At every road they reserve a lane just for you. At every intersection they post a big sign that says for a given lane which way to turn and what new lane to take. MPLS Explicit Routing Multiple Label-Switched Paths (LSPs) between an ingress-egress pair can be efficiently established The Need for Traffic Engineering No automatic load balancing among LSPs Design Goals Distributed load-balancing algorithmNeed no extra network supportMinimal packet reordering requiredGeneral framework for traffic engineeringInternet Draft: draft-widjaja-mpls-mate-02.txt Two-State Adaptive Traffic Engineering Functional Units in Ingress LSRs Probe packets are sent to estimate the relative one-way mean packet delay and packet loss rate along the LSP Traffic Engineering Problem For each Ingress-Egress pair s:InputOffered Load: asSet of LSPs: Ps (an LSP p)OutputVector of traffic splits: ls lsp = as Problem Formulation Define a cost Cp , for an LSP p, as a function of link utilization lspEach ingress-egress pair minimizes the sum of the cost function of each LSP subject to a feasible traffic split Min C(ls)Cp (lsp)s.t.lsp = as, lsp >

0 Understanding the Cost Function Not necessarily a perfect cost functionHelp steer network toward desirable operating pointAllows systematic derivation and refinement of practical traffic engineering schemes Solution Approach Optimality CriterionOptimal if paths with positive flow have minimum (and equal) cost derivativesGradient Projection AlgorithmShift traffic from paths with highest derivatives to paths with lowest derivatives by a small amount each iteration Asynchronous Environment Feedback delays (probe measurements): non-negligible different delays for LSPs time-varyingMany ingress-egress routers shift traffic independently at different times likely with different frequencies Convergence under AsynchronousConditions The algorithm will converge provided the cost function satisfies certain requirementsStarting from any initial rate vector l(0), the limit point of the sequence {l (t)} is optimal, provided the step size is sufficiently smallBound on step size estimates the effect of asynchronism Packet-level Discrete Event Simulator Entities: Packets, Routers, Queues, and LinksSimulated Functional Units Measurement and Analysis Traffic Engineering Assume traffic already filtered into binsBoth Poisson and Long-range dependent traffic (DAR) Experiment Setup Aggregate Utilization on Shared Links Packet Loss on Shared Links Conclusion MPLS Adaptive Traffic Engineering an end-to-end solution without network support distributed load-balancing steer networks toward "optimal" operating point under asynchronous network conditions validated in simulation