Cassie Heart
@cassie
Wake up, babe. New spicy post dropped: https://cassieheart.substack.com/p/crypto-unchained
5 replies
0 recast
0 reaction
Rohit Kulshreshtha
@rohit
Are shards and roll ups the same thing? I can understand roll ups means publishing to the L1 block chain - but do shards necessarily need that? I’m comparing to shards in regular dbs- the only time you need dist txn is if you cross shard boundaries, but if you can guarantee locality, it is avoidable.
1 reply
0 recast
0 reaction
Cassie Heart
@cassie
They’re not the same, but effectually have the same outcome — the current shard proposal has receipts landing on the beacon chain so that they can be used for cross-shard transactions. Guaranteed locality grows implausible as the number of shards increases.
1 reply
0 recast
0 reaction
Rohit Kulshreshtha
@rohit
Thanks for sharing. Are there any proposals to reduce the 12s time? It seems like a big lever.
1 reply
0 recast
0 reaction
Cassie Heart
@cassie
Vitalik has explicitly argued against doing this, as it would result in the network becoming more centralized: https://www.reddit.com/r/ethereum/comments/slzfsd/why_wouldnt_proof_of_stake_drastically_reduce/hvu9ekc/
2 replies
0 recast
0 reaction
Rohit Kulshreshtha
@rohit
Are there any nuggets of wisdom about the 4MB block size or proposals to increase? Granted it doesn’t help for the dApp latency since it only increases bandwidth. The link you shared makes it sound like the work done per block is O(log(block size)): that seems like a nice property for increasing bandwidth.
2 replies
0 recast
0 reaction
Rohit Kulshreshtha
@rohit
Thank you so much for sharing all of this, I learned a tonne of new stuff today. The lower latency connections that Vitalik refers to kinda reminds me of arbitrage techniques in low latency high frequency trading.
0 reply
0 recast
0 reaction
