IPNI: A federated pathfinder to just about any CID
summary
IPNI catalogs and serves content-routing records at planetary scale, ingesting about 4 billion CIDs/day (peaked ~20B/day) and answering lookups in one hop with ~30 ms 90th‑percentile latency, handling ~50,000 requests/min. It uses a federated set of servers (three instances rolling out federation) with chained key‑value clusters and per‑provider chains of advertisements, stores encrypted records via double‑hash encrypted records for reader privacy, integrates with Kubo and Lotus, and runs as an open foundation public good.
AI-written summary · about
chapters (6)
- 0:00Introduction to IPNI
- 3:56The discovery problem
- 7:29Dumb servers and federation
- 10:18IPNI architecture and specs
- 13:11Current performance and stats
- 14:55IPNI foundation and future
related talks
30m🛠 Saturn & IPNI: Finding and fetching Web3 Content
20mTackling Container Networking at Scale
05mLightning Talk: Unveiling 5 Essential Insights from Cilium Implementation in a...
29mThe Swiss Army Knife of Cloud Native Networking
35mNext-level Kubernetes Networking with Cilium - Nico Meisenzahl and Philip Welz
17mMonitoring DNS with Open-Source Solutions
35mCNI Intro
16mTo Infinity and Beyond: Exploring the Secret Worlds of the Internet