cc-pipe: Breaking Systemic Bottlenecks in RPKI Data Supply Chain with Concurrent and...
summary
cc-pipe replaces the RP monolithic validation barrier with a concurrent pipeline that generates local view updates per resource certificate, uses a predictive conflict graph and reference counting to ensure strong consistency, and groups conflicting updates for simultaneous distribution. Implemented as a reusable library and integrated into Routinator and Fort, it cuts average validation latency by up to 73.3% in a month-long live deployment, saves hours under projected future loads, tolerates dozens of slow publication points, and adds only modest RP memory/CPU overhead.
AI-written summary · about
chapters (10)
- 0:00RPKI introduction and challenges
- 3:24Bottlenecks in RPKI data supply
- 5:35Consistency vs pipelining dilemma
- 6:20cc-pipe three-step roadmap
- 12:31Router overhead control in cc-pipe
- 14:59Implementation and evaluation
- 15:22Real-world deployment results
- 16:44Scalability results
- 17:28Robustness against slow publishers
- 18:10Conclusion and Q&A
related talks
21mR2P2: Making RPCs first-class datacenter citizens
14mEmpower Programmable Pipeline for Advanced Stateful Packet Processing
15mFrom Address Blocks to Authorized Prefixes: Redesigning RPKI ROV with a Hierarchical...
05mJump-starting back-office connections
33mAdvanced rp_ filter
14mA large-scale deployment of DCTCP
54mPDCP Packet Processing Library in DPDK - Libpdcp
15mIsolation Mechanisms for High-Speed Packet-Processing Pipelines
