You are viewing a single comment's thread from:

RE: Announcing New Public Full Node

in #steem6 years ago

Great work! In your view @markymark has there been real progress by Stinc towards reducing the RAM requirements of full nodes yet or is it just promises?

Posted using Partiko iOS

Sort:  

Account History on RocksDB greatly reduced full node memory usage by almost half. HiveMind will reduce it further, I am not running HiveMind though, I am doing full SteemD & RocksDB. You still need a lot to run a HiveMind based full node.

You are looking at 4-5 machines rather than one, but they can be each be lower ram than one big one. The more ram you get, the price goes up exponentially not linearly, that's where the savings come in.

Once everything can be stored on disk, then you will really see cheap nodes.

As far as I understand that has already been done.

https://steemit.com/dtube/@exyle/e05ebhbs