AnyPro: Preference-Preserving Anycast Optimization based on Strategic AS-Path Prepending
summary
AnyPro is a system that uses AS-path prepending and an ICMP probing testbed to derive constraint-based ASPP settings that steer clients to preferred PoPs. It discovers ASPP-sensitive clients with max-min polling, formulates a weighted max-satisfaction optimization, and resolves conflicting constraints via iterative binary tightening. Deployed on a global 20 PoP testbed, it drives 90th-percentile RTT down (examples: baseline 271 ms, AnyPro 58 ms), and reports large preference-alignment gains.
AI-written summary · about
chapters (8)
- 0:00AnyPro introduction
- 1:24Anycast deployment challenges
- 4:37AS path prepending mechanism
- 6:25Preference alignment optimization
- 9:14AnyPro probing system
- 13:46AnyPro performance benefits
- 16:16Theory versus real-world RTT
- 17:42ML comparison and summary
related talks
29mUsing the Internet as Your Load-Balancer
20mContra: A Programmable System for Performance aware Routing
17mTowards provably performant congestion control
14mPoseidon: Efficient, Robust, and Practical Datacenter CC via Deployable INT
19mtpprof: A Network Traffic Pattern Profiler
23mAutomatic Traffic Scheduling for Internet Connectivity Failures
21mR2P2: Making RPCs first-class datacenter citizens
20mPostMan: Rapidly Mitigating Bursty Traffic by Offloading Packet Processing