Post #1645534

2026-04-17 16:19 UTC

I'm anticipating some scale issues with the fairly straightforward way I'm storing data. I have ideas towards them, but I'm also realizing that because of its decentralized nature there's entire categories of scale issues sawa just opts out of. getting this to work with thousands of concurrent users running through a single server is technically difficult and also a non-goal. I think an explicit part of sawa's answer to scaling is to make standing up your own server extremely easy, easier than mastodon.

Replies (1)

@nasser@merveilles.town 2026-04-17 16:32

people really resonated with sawa's goal of implementability too. inspired by 100r (https://100r.co/site/weathering_software_winter.html) sawa is committed to being every easy to extend with new storage and network backends. I have it down to 9 verbs, 4 for the key value storage system and 5 for the network. for storage: - get - put - drop - list for network: - hello - have - want - give - thank sawa is entirely driven by it's storage and network layer and as a result entirely derived from these 9 verbs. if you can implement these verbs, you can offer sawa a new storage and/or network backend. the current prototype uses leveldb for storage and socket io for the network. but storage from redis, the filesystem, s3 or whatever are totally doable. networking over meshtastic or local Bluetooth are also totally doable, just implement the verbs.

Open ##1645535