Path Computation Element PCE Architecture draftashpcearchitecture01 txt Jerry

Path Computation Element (PCE) Architecture (draft-ash-pce-architecture-01. txt) Jerry Ash AT&T gash@att. com Adrian Farrell Old Dog Consulting adrian@olddog. co. uk JP Vasseur Cisco Systems, Inc. jpv@cisco. com Outline q PCE architecture summary v provided in backup slides v you read the draft q I-D updates based on comments raised on list q next step: working group draft 1

PCE Architecture Summary (see backup slides attached) terminology assumptions motivation for PCE architecture PCE architectural considerations security & confidentiality PCE evaluation metrics PCE architecture overview v composite PCE v external PCE v multiple PCE path computation with inter-PCE communication q architecture non-goal to specify protocols v protocol extensions will be worked out in other IDs q q q q 2

Composite PCE Node 3

External PCE Node 4

Multiple PCE Path Computation 5

Multiple PCE Path Computation with Inter-PCE Communication 6

I-D Updates Based on Issues Raised on List q PCE should advertise its capabilities, for example v set of constraints it can account for (diversity, SRLGs, optical impairments, wavelength continuity, etc. ) v text added to Section 6. 4 q path computation request include if near-disjoint paths acceptable v text added to Section 6. 6 q TED information can include info from sources other than IGP (e. g. LSP routes, reserved bandwidth, measured traffic volume) v needed to perform LSP re-optimization v needed to reconfigure virtual network topology (VNT) lower layer (e. g. , optical) paths v text added to Section 6. 7 q elaborate on advantages of stateful PCE & pitfalls of using stateful PCE in a distributed PCE environment v text added to Section 6. 8 q evaluation metrics should include TED synchronization speed & impact on the data flows v text added to Section 7 7

I-D Updates Based on Issues Raised on List q identify areas for standardization v added Section 5. 5 “Areas for Standardization” – based on PCE Charter q other editorial changes 8

Next Steps q propose draft as PCE WG draft 9

Backup Slides 10

Terminology q path computation element (PCE) v entity (component, application or network node) capable of computing a network path based on network graph & computational constraints v e. g. , PCE computes path of a TE LSP by using TED & bandwidth/other constraints q path computation client (PCC) v any client application requesting a path computation by the PCE q domain v any collection of network elements within a common sphere of address management or path computational responsibility v e. g. , IGP areas, AS, multiple ASs within a SP network, multiple ASs across multiple SP networks q single PCE path computation: single PCE computes a path in a domain q multiple PCE path computation: multiple PCEs compute a path in a domain q centralized computation model: all paths in a domain computed by a single, centralized PCE q distributed computation model: computation of paths in a domain 11 shared among multiple PCEs

Assumptions q PCE may or may not be located at head-end v e. g. nodes on path contribute to path computation (e. g. , loose hops) making them PCEs v path computation may be made by PCE physically distinct from the computed path q path computed by PCE may be v complete: full explicit path of strict hops v partial: mix of strict & loose hops (may be an abstract node such as an AS) q PCE path computation can be used in conjunction with other path computation models v e. g. , inter-AS TE LSP may be computed using PCE in some domains but not others q no assumptions made about PCE implementation v e. g. , could be implemented on a router, LSR, dedicated network server, etc. v PCE function independent of forwarding capability of node on which it is implemente 12

Motivation for PCE Architecture q inter-area/AS optimal path computation (node has partial visibility) q computation of inter-area/AS diverse path (node has partial visibility) q CPU-intensive path computation/global optimization q backup path computation for bandwidth protection with backup capacity optimization q multi-layer networks e. g. TDM network provides connectivity for client-layer (IP, MPLS, L 2, etc. ) q absence of TED or use of non-TE-enabled IGP q node outside routing domain (e. g. , CE to PE path computation) q network element lacks control plan or routing capability 13

PCE Architectural Considerations q synchronization v non-synchronized (e. g. , PCE makes multiple individual path computations to generate set of paths) v synchronized (e. g. , single PCE invokes computations by other PCEs before supplying result to PCC q PCE discovery & load balancing q detecting PCE liveness q PCC-PCE & PCE-PCE communication q PCE TED synchronization q stateful vs. stateless PCEs q monitoring q policy & confidentiality v must preserve confidentiality across multiple SPs v must ensure confidentiality & security of PCC-PCE & PCEPCE messages 14

Security & Confidentiality q PCC-PCE communication v subject to "usual" security issues v snooping not a significant issue – might want to encrypt v spoofing is very serious – must offer strong authentication – protocol is P 2 P so this is relatively easy v Do. S important because of 'centralized' nature of PCE q PCE-PCE communication v same as for PCC-PCE, but add confidentiality q confidentiality (protection of domain topology information) v use loose routes v PCE encrypts ERO segments – decrypt on entry to domain v replace ERO segment with cookie – entry point to domain consults local PCE using cookie to retrieve next ERO segment 15

PCE Evaluation Metrics q q q q optimality scalability load sharing multiple path computation reoptimization path computation time network stability synchronization v between TED & network topology/resource states v speed of TED synchronization v impact of synchronization on data flows 16