Our great sponsors
-
-
-
WorkOS
The modern identity platform for B2B SaaS. The APIs are flexible and easy-to-use, supporting authentication, user identity, and complex enterprise features like SSO and SCIM provisioning.
2) Improved throughput. There are a few different algorithms that have been developed by Jito Labs and Solana Labs to improve the amount of parallelism happening during block packing (when a leader validator builds a block from transactions) and replay (when other validators replay blocks to make sure they're good). Solana validators have super good hardware but we don't even use that many of the cpu threads for actually executing transactions, and the current block space limits (48 million compute units) were conceived for the use of 4 cpu threads (our validator server has 48 cpu threads). The older block packing algorithm was also really bad at handling situations where a certain transaction type was spammed heavily (nft mints for example) and degraded parallelism. Anyway, it seems that these newer replay algorithms are over 2x better than the current implementations -- you can find more about them in the depths of github or discord. Some seem close to production-ready and I think we will see faster confirmation times and increasing of the block compute limits. ABIv2 is another architectural change that will substantially improve throughput as well (afaik) but I know less on this subject. I'm not sure when this will be ready but I've heard that the earlier half of the year is the target https://github.com/solana-labs/solana/issues/27384. There are other upcoming changes to look forward to, but it's probably best to summarize in another post.
3) Jump Trading/Crypto has an incredible engineering team (might be the best in crypto tbh) and is working on the second Solana validator implementation, Firedancer, which should make the network even more performant and robust. They will be going through the current validator implementation by Solana Labs and increasing documentation around it and presumably finding areas for improvement of the existing one as well. Kevin Bowers is leading this project and is the Chief Science Officer of Jump Trading and has an incredible amount of experience in high performance computing https://www.linkedin.com/in/kevinjbowers. He has been committing code on weekends and you can track things on the Firedancer github. https://github.com/firedancer-io/firedancer