From 594a1ecf13ab5665e749b9adaab168fa77020a25 Mon Sep 17 00:00:00 2001 From: Sasha Date: Wed, 3 Jul 2024 01:11:05 +0200 Subject: [PATCH] improve service node pool sub section --- informational/req-res-reliability.md | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/informational/req-res-reliability.md b/informational/req-res-reliability.md index 3470efb..142d144 100644 --- a/informational/req-res-reliability.md +++ b/informational/req-res-reliability.md @@ -42,9 +42,8 @@ To address this we suggest following metrics: ### Peers and connection management -- Light node should retain a pool of reliable service nodes per each protocol. -In case a protocol failed to use any service node more than once - -connection to it should be dropped and a new service node should be connected and added to the pool instead. +- Light node should maintain a pool of reliable service nodes for each protocol. +In case service node fails to serve protocol request from a light node 3 times - light node should drop connection to it and a new service node should be connected and added to the pool instead. - During discovery of new service nodes it is better to filter out based on ENR / multiaddress. For example in some cases `circuit-relay` addresses are not needed when we try to find and connect to service nodes directly.