~/talk

The Eternal Tussle: Exploring the Role of Centralization in IPFS

Talkat 21st USENIX Symposium on Networked Systems Design and Implementation (NSDI '24) · 2024-04-16..04-18· [peer-to-peer]

speakers

links:none on file

name:Dennis Trautwein

x:dtrautwein_eu

name:Yiannis Psaras

website:https://probelab.io

name:Ignacio Castro

links:none on file

name:Will Scott

links:none on file

name:Aravindh Raman

links:none on file

name:Gareth Tyson

links:none on file

resources

video:https://youtu.be/mJ0OCrWD5YY

slides:none on file

summary

IPFS adds three centralized components and measures concrete gains and costs: IPNI indexer accepts massive publication and holds ~100× more provider records than the DHT (top providers 100–1,000×), with seven federated operators to mitigate its single‑point‑of‑failure risk; Hydra Boosters run thousands of virtual DHT heads and a shared DB, cutting median lookup time by 3.7–36.5% but costing thousands of dollars per month; HTTP Gateways cache ~80% of requests, making cached fetches ~100× faster while sacrificing end‑to‑end content verification.

AI-written summary · about

chapters (8)

related talks