Closed weboko closed 2 months ago
Draft for now before https://github.com/waku-org/js-waku/pull/1958 is merged.
Path | Size | Loading time (3g) | Running time (snapdragon) | Total time |
---|---|---|---|---|
Waku node | 181.09 KB (+0.16% 🔺) | 3.7 s (+0.16% 🔺) | 1.4 s (-17.96% 🔽) | 5 s |
Waku Simple Light Node | 181.03 KB (-0.09% 🔽) | 3.7 s (-0.09% 🔽) | 1.7 s (+9.31% 🔺) | 5.3 s |
ECIES encryption | 23.08 KB (0%) | 462 ms (0%) | 448 ms (+181.83% 🔺) | 909 ms |
Symmetric encryption | 22.55 KB (0%) | 452 ms (0%) | 330 ms (-25.77% 🔽) | 781 ms |
DNS discovery | 72.42 KB (0%) | 1.5 s (0%) | 837 ms (+3% 🔺) | 2.3 s |
Peer Exchange discovery | 74.1 KB (0%) | 1.5 s (0%) | 1.1 s (+4% 🔺) | 2.5 s |
Local Peer Cache Discovery | 67.64 KB (0%) | 1.4 s (0%) | 657 ms (-31.03% 🔽) | 2.1 s |
Privacy preserving protocols | 38.87 KB (0%) | 778 ms (0%) | 496 ms (-7.34% 🔽) | 1.3 s |
Waku Filter | 111.76 KB (+0.23% 🔺) | 2.3 s (+0.23% 🔺) | 1.2 s (+62.87% 🔺) | 3.4 s |
Waku LightPush | 110.11 KB (0%) | 2.3 s (0%) | 938 ms (-5.49% 🔽) | 3.2 s |
History retrieval protocols | 110.72 KB (0%) | 2.3 s (0%) | 828 ms (+0.55% 🔺) | 3.1 s |
Deterministic Message Hashing | 7.29 KB (0%) | 146 ms (0%) | 77 ms (+17.66% 🔺) | 223 ms |
Why can't we use the existing keep alive manager for this? Or why is it insufficient?
Why can't we use the existing keep alive manager for this? Or why is it insufficient?
This is because keepAliveManager
does libp2p pings
Overall something like that can be implemented for filter pings
once logic becomes more sophisticated: perhaps if next steps are implemented from here https://github.com/waku-org/js-waku/issues/1923#issuecomment-2070939508
Problem
Filter doesn't keep keep subscription forever and requires periodic pings.
Solution
Implement
keep alive
feature to do automatic pings to the peers.Notes