From f44584e61a3ad40d9e24cbbd85ea9f9df98e220a Mon Sep 17 00:00:00 2001 From: Thane Thomson Date: Thu, 15 Sep 2022 11:05:42 -0400 Subject: [PATCH] Add more pros/cons from discussion Signed-off-by: Thane Thomson --- docs/architecture/adr-082-data-companion-api.md | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/docs/architecture/adr-082-data-companion-api.md b/docs/architecture/adr-082-data-companion-api.md index 15ee52a13..0431908a3 100644 --- a/docs/architecture/adr-082-data-companion-api.md +++ b/docs/architecture/adr-082-data-companion-api.md @@ -314,12 +314,17 @@ in a patch release). - Paves the way toward greater architectural simplification of Tendermint so it can focus on its core duty, consensus, while still facilitating existing use cases. +- Allows us to eventually separate out non-operator-focused RPC functionality + from Tendermint entirely, allowing the RPC to scale independently of the + consensus engine. - Can be rolled out as experimental and opt-in in a non-breaking way. - The broad nature of what the API publishes lends itself to reasonable long-term stability. ### Negative +- Keeping existing Tendermint functionality would involve operators having to + run an additional process, which increases operational complexity. - It is unclear at present as to the impact of the requirement to publish large quantities of block/result data on the speed of block execution. This should be quantified in production networks as soon as this feature can be rolled out