Stellar core developer discusses blockchain scalability through commutative APIs and parallel execution, presenting research on batch exchange markets and machine code smart contracts to enable orders of magnitude more on-chain computation without sacrificing decentralization.
A Stellar core developer presents research on achieving blockchain scalability beyond single-CPU limitations. The talk covers why current parallel execution approaches fail with conflicting transactions, introduces the Scalable Commutativity Rule from operating systems research, and demonstrates how batch exchange markets (inspired by Arrow-Debreu economics) make trading operations commutative and thus parallelizable to 250,000+ trades per second. The speaker also discusses Groundhog, a prototype blockchain achieving 800,000 payments per second with 192 threads, and presents research on executing raw machine code smart contracts for better performance and developer equity. The core argument is that eliminating computational scarcity on-chain will unlock innovation and enable competition with traditional finance.