ift-ts:dst:ift:2026q2-libp2p-evaluation
Description
Test libp2p on each new version or requested feature and look for regressions, learn scaling properties and run scaling studies, understand the limits of Waku and its behaviour. Deliver reports and actionable insights. Do this monthly, reliably, with documentation of findings.
The scope of this commitment depends on the P2P team work and improvements, and it is subjected to change.
Task list
Schedule note: Dates reflect quarter bounds; update when actual timing is known.
Regression testing (recurring)
- fully qualified name:
ift-ts:dst:ift:2026q2-libp2p-evaluation:regression-testing - owner: Pearson
- status: not started
- start-date: 2026/04/01
- end-date: 2026/06/30
Description
Run different scenarios and collect evidence and data of libp2p’s behaviour.
Test for known regressions that have occurred in the past and ensure they don’t happen again.
Deliverables
- Code:
- Reports:
nim-lsquic benchmark
- fully qualified name:
ift-ts:dst:ift:2026q2-libp2p-evaluation:nim-lsquic-benchmark - owner: Mamoutou
- status: in progress (70%)
- start-date: 2026/04/01
- end-date: 2026/06/30
Description
Run the core part of quic in nimlibp2p in an isolated manner, instead running the dst-test-node. We want to perform similar scenarios than the ones we are doing for regression experiments. This means to set up a kubernetes deployment, and make the new scripts to have a similar behavior than the dst-test-node in terms of connections and throughput. We need to study if the latencies reported by quic in the nim-libp2p regression experiments are caused by this implementation or if there is another factor.