Snapchain is getting close to an alpha release! 

Snapchain is a new blockchain-like system that stores your casts, reactions and follows and keeps them in sync between apps. 

It's designed to be faster and scale better than the current hub network. 

https://github.com/farcasterxyz/protocol/discussions/207

Technowatermelon. Elder Millenial. Building Farcaster. 

nf.td/varun

"Faster" sync comes from putting all transactions into blocks. 

If a node gets disconnected for a bit, it can just look up how many blocks it missed and download them. This simply isn't possible today with the hub and deltagraph system, which is why there are sometimes long delays or missed messages between apps.

"Scale" comes from sharding. 

There's a pretty neat sharding design inspired by NEAR's Nightshade that lets us split up accounts into different shards. So we can easily get 10 and maybe even a 100 machines producing blocks simultaneously, which is much higher than most blockhains. 

Our target is to cross 9000 TPS with this design, which should support 2M+ daily users.

Snapchain should be in alpha later this month. 

It will be "feature complete" and on parity with existing hubs, and we plan to launch the testnet at the end of this month or early next month. 

If all goes well, we'll be fully on Snapchain by the end of January.

Snapchain should be in alpha later this month. 

It will be "feature complete" and on parity with existing hubs, and we plan to launch the testnet at the end of this month or early next month. 

If all goes well, we'll be fully on Snapchain by the end of January. 

https://github.com/farcasterxyz/snapchain-v0

Another nice side effect of using blocks is that rate limits won't break sync. 

Every account has a rate limit of how many messages per second it can post. This is high enough that a human wouldn't trigger it, but bots sometimes do. 

Rate limits on hubs have to be local because there's no way to order messages across hubs. So one hub might reject your messages because you hit it too fast, while another might say its ok. 

This leads to some weird outcomes where hubs get out of sync and it makes it even hard to track down other sync issues.