What do people really mean when they talk about blockchain performance? Aptos Labs' Head of Research Sasha Spiegelman joins a16z crypto Head of Research Tim Roughgarden and show host Robert Hackett to discuss ways technologists are trying to maximize throughput and minimize latency for blockchains.
with @SashaSpiegelman @Tim_Roughgarden @rhhackett
Welcome to web3 with a16z. Today we’re talking about the ins and outs of blockchain performance.
How does the “speed” of one chain compare to another? What are the tricky ways that people talk about important metrics like throughput and latency? And how do design choices across things like consensus and execution result in performance tradeoffs?
Our guest today is Aptos Labs’ Head of Research Sasha Spiegelman, who is helping build out and develop technologies he originally co-created at Meta. We’re also joined by a16z crypto Head of Research and Columbia University professor Tim Roughgarden, who specializes in the intersection of computer science and economics, which is fitting as we dig into technical details throughout.
Timestamps:
00:00 introduction
02:48 understanding blockchain metrics
07:59 consensus protocols: PBFT and beyond
14:56 innovations from Meta’s Novi to Aptos
20:51 optimizing blockchain execution (with BlockSTM)
25:23 tech debate: dynamic vs. static parallelism
27:55 open source and the prisoner’s dilemma
29:15 conclusion
***
As a reminder, none of the content should be taken as tax, business, legal, or investment advice. Please see a16z.com/disclosures for more important information, including a link to a list of our investments.