ViAggre: Making Routers Last Longer! Hitesh Ballani Paul Francis, Tuan Cao and Jia Wang Cornell University and AT&T Labs – Research HotNets 2008
Motivation: Rapid Routing Table Growth 300000 300000 282,000 prefixes 282,000 prefixes Active BGP entries (FIB Size) Active BGP entries (FIB Size) (Sep’08) (Sep’08) 250000 250000 200000 200000 150000 150000 100000 100000 50000 50000 0 0 88 88 90 90 92 92 94 94 96 96 98 98 00 00 02 02 04 04 06 06 08 08 Year Year
Motivation: Rapid Routing Table Growth 500000 500000 ?? Active BGP entries (FIB Size) Active BGP entries (FIB Size) Rapid future growth 400000 400000 300000 300000 ◮ IPv4 exhaustion 200000 200000 ◮ IPv6 deployment 100000 100000 0 0 88 88 90 90 92 92 94 94 96 96 98 98 00 00 02 02 04 04 06 06 08 08 14 14 Year Year
Typical Router Innards Route Processor CPU RIB Line Card Line Line Line Card Card Card FIB ASIC
Typical Router Innards Route Processor Routing Information Base CPU RIB (DRAM $) Line Card Line Line Line Card Card Card FIB ASIC
Typical Router Innards Route Processor CPU RIB Line Card Line Line Line Card Card Card FIB ASIC Forwarding Information Base (SRAM $$$)
Does (FIB) Size Matter? Technical concerns ◮ More Memory ◮ More Processing ◮ Power and Heat dissipation problems
Does (FIB) Size Matter? Technical concerns ◮ More Memory ◮ More Processing ◮ Power and Heat dissipation problems Business concerns ◮ Less cost-effective networks ◮ Price per byte forwarded increases ◮ Router memory upgrades
Does (FIB) Size Matter? Technical concerns ◮ More Memory ◮ More Processing ◮ Power and Heat dissipation problems Business concerns ◮ Less cost-effective networks ◮ Price per byte forwarded increases ◮ Router memory upgrades ISPs are willing to undergo some pain to reduce FIB size
Routing Scalability Problem Space [Deering, ’96] [O’Dell, ID’97] [Zhang et. al., ICNP’06] FIB growth [Farinacci, ID’07] RIB growth [Massey et. al., ID’07] [Jen et. al., HotNets’08] [Francis, CNIS’94] Routing Convergence, [Deering et. al., ID’00] Update Churn, .... [Hain, ID’02] [Krioukov et. al., Arxiv’05]
Routing Scalability Problem Space [Deering, ’96] [O’Dell, ID’97] [Zhang et. al., ICNP’06] FIB growth [Farinacci, ID’07] RIB growth [Massey et. al., ID’07] [Jen et. al., HotNets’08] [Francis, CNIS’94] Routing Convergence, [Deering et. al., ID’00] Update Churn, .... [Hain, ID’02] [Krioukov et. al., Arxiv’05] Separate edge from the core
Routing Scalability Problem Space [Deering, ’96] [O’Dell, ID’97] [Zhang et. al., ICNP’06] FIB growth [Farinacci, ID’07] RIB growth [Massey et. al., ID’07] [Jen et. al., HotNets’08] [Francis, CNIS’94] Routing Convergence, [Deering et. al., ID’00] Update Churn, .... [Hain, ID’02] [Krioukov et. al., Arxiv’05] Geographical routing
Routing Scalability Problem Space [Deering, ’96] [O’Dell, ID’97] [Zhang et. al., ICNP’06] FIB growth [Farinacci, ID’07] RIB growth [Massey et. al., ID’07] [Jen et. al., HotNets’08] [Francis, CNIS’94] Routing Convergence, [Deering et. al., ID’00] Update Churn, .... [Hain, ID’02] [Krioukov et. al., Arxiv’05] Compact routing
Routing Scalability Problem Space [Deering, ’96] [O’Dell, ID’97] [Zhang et. al., ICNP’06] FIB growth [Farinacci, ID’07] RIB growth [Massey et. al., ID’07] [Jen et. al., HotNets’08] [Francis, CNIS’94] Routing Convergence, [Deering et. al., ID’00] Update Churn, .... [Hain, ID’02] [Krioukov et. al., Arxiv’05] All require architectural change So many ideas, so little impact!
Routing Scalability Problem Space [Deering, ’96] [O’Dell, ID’97] [Zhang et. al., ICNP’06] FIB growth [Farinacci, ID’07] RIB growth [Massey et. al., ID’07] [Jen et. al., HotNets’08] [Francis, CNIS’94] Routing Convergence, [Deering et. al., ID’00] Update Churn, .... [Hain, ID’02] [Krioukov et. al., Arxiv’05] Tackle routing scalability through a series of incremental, individually cost-effective upgrades
Routing Scalability Problem Space [Deering, ’96] [O’Dell, ID’97] [Zhang et. al., ICNP’06] FIB growth [Farinacci, ID’07] RIB growth [Massey et. al., ID’07] [Jen et. al., HotNets’08] [Francis, CNIS’94] Routing Convergence, [Deering et. al., ID’00] Update Churn, .... [Hain, ID’02] [Krioukov et. al., Arxiv’05] This Paper: Focuses on reducing FIB size
Virtual Aggregation, aka ViAggre A “configuration-only” approach to shrinking router FIBs ◮ Applies to legacy routers ◮ Can be adopted independently by any ISP Key Insight: Divide the routing burden A router only needs to keep routes for a fraction of the address space
Talk Outline ◮ Introduction[]y ◮ ViAggre: Basic Idea[]y ◮ ViAggre Design[]y ◮ Evaluation[]y ◮ Deployment[]y ◮ Conclusions
ViAggre: Basic Idea 0.0.0.0 IPv4 Address Space PoP C PoP A 255.255.255.255 PoP B External External Router Router ISP Today: All routers have routes to all destinations
ViAggre: Basic Idea 0.0.0.0 Virtual 0/2 Prefixes 64/2 128/2 PoP C PoP A 192/2 255.255.255.255 PoP B External External Router Router ISP Divide address space into Virtual Prefixes (VPs)
ViAggre: Basic Idea 0.0.0.0 Virtual 0/2 Prefixes 64/2 128/2 192/2 255.255.255.255 Aggregation Points for Green VP External External Router Router Assign Virtual Prefixes to the routers Routers only have routes to a fraction of the address space
ViAggre: Basic Idea 0.0.0.0 Virtual 0/2 Prefixes 64/2 128/2 192/2 255.255.255.255 Aggregation Points for Green VP External External Router Router How to achieve such division of the routing table? Without changes to routers and routing protocols Without cooperation from external networks
Talk Outline ◮ Introduction[]y ◮ ViAggre: Basic Idea[]y ◮ ViAggre Design[]y ◮ Evaluation[]y ◮ Deployment[]y ◮ Conclusions
ViAggre Control-Plane 0.0.0.0 0/2 64/2 128/2 192/2 Full Routing Table 255.255.255.255 External External Router Router eBGP Peers may advertise full routing table
ViAggre Control-Plane 0.0.0.0 0/2 64/2 128/2 192/2 Full Routing Table 255.255.255.255 External External Router Router Load full routing table into RIB Supress all but blue routes from FIB FIB Suppression Blue routers only load blue routes into their FIB
Data-Plane paths 0.0.0.0 0/2 64/2 128/2 Packets destined to a prefix in Red 192/2 Virtual Prefix 255.255.255.255 External External Router Router Consider packets destined to a prefix in the red VP
Data-Plane paths 0.0.0.0 0/2 64/2 A 128/2 1 192/2 2 I A2 255.255.255.255 E External X External Router Router ViAggre path Ingress (I) → Aggregation Pt (A) → Egress (E)
Ingress → Aggregation Point 0.0.0.0 0/2 64/2 A 128/2 1 192/2 I A2 255.255.255.255 E External X External Router Router Router I doesn’t have a route for destination prefix
Ingress → Aggregation Point 0.0.0.0 0/2 Advertise 64/2 A Red VP 128/2 1 192/2 I A2 255.255.255.255 E External X External Router Router Aggregation Points advertise corresponding Virtual Prefixes
Ingress → Aggregation Point 0.0.0.0 0/2 Advertise 64/2 A Red VP 128/2 1 192/2 I A2 255.255.255.255 Prefix Next-Hop E External .... X P1 External Router .... P2 Router A 0/2 .... 128/2 .... 192/2 Blue router has a route for the red Virtual Prefix
Aggregation Point → Egress 0.0.0.0 Prefix Next-Hop P3 X 0/2 .... P4 .... 64/2 64/2 A .... 128/2 .... 128/2 192/2 192/2 2 I A2 255.255.255.255 E External X External Router Router Aggregation Pt. A tunnels packet to external router
Aggregation Point → Egress 0.0.0.0 0/2 64/2 A 128/2 1 192/2 I A2 255.255.255.255 E External X External Strip tunnel header Router Router from outgoing pkts Egress Router strips the tunnel header off outgoing packets
Failure of Aggregation Point 0.0.0.0 0/2 64/2 A 128/2 1 192/2 I A2 255.255.255.255 E External X External Router Router What if Aggregation Pt. A fails?
Failure of Aggregation Point 0.0.0.0 0/2 64/2 A 128/2 192/2 I A2 255.255.255.255 Prefix Next-Hop E External .... X P1 External Router .... P2 Router 0/2 A2 .... 128/2 .... 192/2 Router I installs the route advertised by A2
Failure of Aggregation Point 0.0.0.0 0/2 64/2 A 128/2 192/2 I A2 255.255.255.255 Prefix Next-Hop E External .... X P1 External Router .... P2 Router 0/2 A2 .... 128/2 .... 192/2 Packets are re-routed appropriately
ViAggre’s impact on ISP’s traffic 0.0.0.0 0/2 64/2 A 128/2 1 192/2 I A2 255.255.255.255 E External X External Router Router ViAggre paths can be longer than native paths Traffic stretch, increased router and link load, etc.
Popular Prefixes Traffic volume follows power-law distribution ◮ 95% of the traffic goes to 5% of prefixes ◮ Has held up for years Install “Popular Prefixes” in routers ◮ Stable over weeks ◮ Mitigates ViAggre’s impact on the ISP’s traffic
Recommend
More recommend