Evolution Towards Global Routing Scalability draft-zhang-evolution-01 Varun Khare [email protected] Beichuan Zhang [email protected] Lixia Zhang [email protected] IETF74, March 09 1 Internet Routing Scalability: a problem? • DFZ routing table is growing uncontrollably • Different parts of the Internet view the scalability problem differently. – Most Stub ASes don’t carry full table internally – Certain ISPs can afford to upgrade routers – Within AS some routers experience problems more severely • No unanimous agreement about the problem 2 Internet Routing Scalability: a problem? • Growth in DFZ routing table implies following for routers: – FIB size growth – RIB size growth – Increase in Churn due to Edge Dynamics 3 Controlling FIB Size on Old Routers • Solution: Virtual Aggregation deployable by single ISP [ViAggre by Ballani et. al., HotNets’ 08] – Don’t need coordination with anyone else – Only need to make change in router configurations – No impact upon routing operations of other networks • ViAggre brings immediate FIB reduction 4 Virtual Aggregation is Map-Encap • Map destination prefix to local exit router • APR holds Map of detailed prefixes • APR encaps packet to BR since other routers don’t know how to reach prefix 1.2/16 AS1 APR1 1/8 decap encap 1/8 APR1 PE3 PE2 PE3 PE1 5 Inter-AS overhead of VA • If neighboring ASes deploy VA then packet can experience stretch and encap-decap cost in both ASes – Both AS1 and AS2 are mapping destination prefix to their local exit router 1.2/16 PE3 AS1 encap encap 1.2/16 CE1 decap APR1 AS2 PE4 APR2 decap PE2 PE1 1/8 stretch APR1 PE3 1/8 stretch APR2 CE1 DST 1.2/16 6 Solve Inter-AS overhead of VA • Propagate mapping info across ASes Inter-AS ViAggre [Xiaohu talk IETF74 RTGW] – Only need to Map-encap once – Packet can cross ASes not keeping detailed FIB – Piggyback mapping info on BGP updates 1.2/16 PE4 encap PE4 PE2 1.2/16 PE4 APR2 AS1 AS2 PE4 APR1 PE2 PE3 PE1 decap 1/8 1/8 APR1 APR2 CE1 DST 1.2/16 7 RIB size problem • After Inter-AS ViAggre – Non-APR routers store full Table with Map Info – APR store routes to virtualized prefixes with Map Info • Mapping Table overhead – Same mapping info can be announced through different AS paths – Routers end up storing multiple copies of map info • As more ISPs share mapping info, the mapping table will become very big. 8 Solutions to RIB size problem RIB Size Problem Proposed Solution At Internal Routers No need to stores routes to virtualized prefixes BRs filter announcement of virtualized prefixes At Border Routers (BR) No need to store full routing table APRs exchange BGP updates with map info on separate BGP session with APRs in neighboring ISPs via multi-hop BGP session At APRs No need to store multiple copies of map info Upgrade APRs BGP implementation 10 A side note: whose prefixes get aggregated out? • Each ISP will keep its internal routes • ISPs will exchange reach ability of their networks – So that an ISP can encapsulate packets from ingress routers to (possibly another ISP’s) egress routers that connect to destination user sites • An ISP aggregates out the prefixes that belong to remote (non-direct) users 11 Excessive Churn due to edge dynamics • APRs share the edge churn load and shield other non-APR routers from it • When border (between provider and customer) failure happens – Mapping updates propagated out to ensure best path taken by data packets. – Mapping updates not propagated out and as data packet reaches failed link, provider network handles it by APT’s data-driven redirection. • Need to upgrade APRs/routers. 12 Evolution towards Scalable Routing • VA, LISP and APT are all Map-Encap schemes • Propose solution to individual problems rather than a comprehensive solution • ISPs decide when to take actions for which problem • As ISPs solve individual problems, the routing architecture is steered to scalable routing • This talk: to illustrate feasibility of convergence – Not about virtual-aggregation, but take it as first step in the convergence process 15 Evolution –vs– Incremental Deployment • New architectural solutions (like LISP, APT) see big benefits when deployed by lots of ISPs • Incremental deployment of a new design allow an ISP running new architecture to inter-operate with legacy ISPs, but – Deploying a new design means relatively high cost – Immediate gain can be low • In evolutionary path, individual ISPs take actions on their own to solve an immediate problem – Cumulative changes over time can evolve the system to converge towards desired direction 16 Thank You Questions? Comments? 17
© Copyright 2025 Paperzz