[Update] Week #50 - peers, sync

This is a bi-weekly update about what I’ve been working on since the last update.

Peers management issue

Related issue: Issue #3017
Related PR: PR #3067

A node keeps too many peers in its state, most of which are dead anyway. The related PR tweaks some params to get rid of the peers quicker. We have a p2p network simulator to simulate how the gossip behaves, but it doesn’t simulate nodes that go offline, restart, etc. We might need to lower the limits on how many nodes are kept in the state, lower the number of peers that are propagated using ping messages, etc (the simulator is useful in these cases).

Throughput test for microblocks

Related PR: PR #3059

The test gives a rough estimates of how long it takes to insert a block with 1 spend transactions into the database including all validations that are done in conductor’s context. These numbers will be useful when these validations are moved outside of the conductor process.

Block/tx validations outside of conductor process

Related issues: Issue #3028, Issue #2167
Related branch: GH-2167-slow-sync

This is still work in progress, not much has been done since the last update.

1 Like