Using Network Virtualization Techniques for Scalable Routing Nick Feamster, Georgia Tech Lixin Gao, UMass Amherst Jennifer Rexford, Princeton University

2 Virtualization for Routing Path Splicing (Georgia Tech) –Run multiple instances of routing protocol –End system signals which instance to use at each hop –Exponential diversity gain, modest complexity Underlay Fused with Overlays (Princeton/GT) –Support overlay functions in routers –Efficient forwarding and scalable monitoring HORN: Hybrid Routing for Overlay Networks (UMass Amherst) –Different nodes see different detailed subgraph –Availability of link state with good scalability

3 Multipath: Promise and Problems Bad: If any link fails on both paths, s is disconnected from t Want: End systems remain connected unless the underlying graph is disconnected ts

4 Path Splicing: Main Idea Step 1: Run multiple instances of the routing protocol, each with slightly perturbed versions of the configuration Step 2: Allow traffic to switch between instances at any node in the protocol t s Compute multiple forwarding trees per destination. Allow packets to switch slices midstream. Feamster, Motiwala, and Vempala, Path Splicing with Network Slicing

5 Perturbations Goal: Each instance provides different paths Mechanism: Each edge is given a weight that is a slightly perturbed version of the original weight –Two schemes: Uniform and degree-based ts Base Graph ts Perturbed Graph

6 Network Slicing Goal: Allow multiple instances to co-exist Mechanism: Virtual forwarding tables a t c s b t a t c Slice 1 Slice 2 dstnext-hop

7 Path Splicing in Practice Packet has shim header with routing bits Routers use lg(k) bits to index forwarding tables –Shift bits after inspection –Incremental deployment is trivial –Persistent loops cannot occur To access different (or multiple) paths, end systems simply change the forwarding bits

8 Reliability Approaches that of Underlying Graph GEANT (Real) and Sprint (Rocketfuel) topologies 1,000 trials p indicates probability edge was removed from base graph Reliability approaches optimal Average stretch is only 1.3 GEANT topology, degree-based perturbations

9 Open Questions and Deployment Can end hosts react quickly enough to recover? –How does the end system find the alternate path? How does splicing perform for other topologies? Deployment Paths –VINI –Overlay –Wireless –Software Router (e.g., Quagga)

10 Splicing: Possible Applications Fast recovery from poorly performing paths Convergence-free routing Data transfer Security: Consistency checking Spatial diversity in wireless networks

11 Path Splicing: High Points Simple: Opaque routing bits provide access to different paths through the network Scalable: Exponential increase in available paths, linear increase in state Stable: Fast recovery does not require fast routing protocols No modifications to existing routing protocols

12 Underlay Fused with Overlays Main idea: Layered routing architecture –Supporting overlay functions on routers –Blur boundary between overlays and underlays Efficient forwarding –Overlay forwarding on line cards –Hosting the overlay control plane Scalable monitoring –Registration of overlay links –Notification of network events Zhu, Rexford, Feamster, Bavier, UFO: A Resilient Layered Routing Architecture

13 Efficient Forwarding Problem: traffic must traverse bottleneck link both inbound and outbound Solution: reflection points in routers Upstream ISP

14 Forwarding on Router Line Cards

15 Scalable Monitoring Notification preserves overlay link abstractions –Message: (overlay source, overlay destination, event) –Routers store state by explicit overlay registration Notification: events that affect performance of overlay applications –Physical failures of routers or links –Reachability failures: withdrawal, session failure –Network congestion

16 Notification of Network Events 1 A 23 4 B C (A,B) (A,C) (A,B) (A,C) (A,B) (A,C) Register for unidirectional overlay links –A->B –B->C

17 Link state routing –Not scalable –A virtual network could be as large as the Internet Distance vector routing –Convergence delay –Scales better Idea: A tunable routing protocol? –Trade scalability for better availability –Each slice runs a fixed protocol with tunable parameter HORN: Scalability vs. Convergence

18 A knows a partial topology (adjacent sub-graph) via link state protocol, and learns routes from the border nodes of the sub-graph. C-H E-G-H D-G-H A-B A-B-C A-D A-D-E A-B-CC-H + A-B-C-H HORN: Hybrid rOuting for oveRlay Networks

19 Benefits and Challenges Benefits: Availability –Tailor to specific topology –Accommodate underlying network constraints –limiting scope of failure notification, geographical proximity Challenges –Potential stretch on path –Expose mapping between overlay and underlay

20 Summary and Question Network virtualization to cheat on scalability tradeoffs –Path diversity vs. scalability –Efficiency vs. scalability –Convergence vs. scalability What are the common abstractions, functions, etc. that the substrate should provide? –Slicing –Nesting –Knobs for granularity control –…?