Sui’s Mysten Labs Technology Prototype Provides First Evidence of Elastic Blockchain Scaling – Blockchain News, Opinion, TV & Careers

adminMarch 21, 2024

Palo Alto, CA, March 20, 2024, Chainwire

Pilotfish, a prototype Sui extension, was able to increase smart contract execution throughput in direct proportion to the number of machines used.

Mysten Labs, a team of leading distributed systems, programming language, and cryptography experts and originators of the Sui Network, today announced a milestone in scaling blockchain capacity. During testing and development in the Sui blockchain environment, Pilotfish, a Sui scaling prototype, successfully demonstrated linear scalability by increasing throughput by 8x with the support of eight systems. Notably, as more machines were added during testing, the latency per transaction decreased, demonstrating the feasibility of linear horizontal scaling for low-latency blockchain transactions for the first time on any blockchain.

This proof-of-concept was officially released as an academic preprint titled “Pilotfish: Decentralized Transaction Execution for Lazy Blockchains” and was initially shared as an article on the open access archive arXiv on January 29, 2024.

“Pilotfish is a game changer as the blockchain industry seeks to achieve the same dynamic capacity growth that is currently only possible through centralized server technologies,” said Lefteris Kokoris-Kogias, one of the paper’s authors and a researcher at Mysten Labs. “Pilotfish has the potential to increase throughput beyond what is currently at the forefront of high-performance blockchains, enabling levels of transaction throughput that no other blockchain in existence can achieve.”

Sui leads the industry in throughput by leveraging multiple cores for parallel transaction execution within individual validators. Using this approach, within a few months of mainnet, Sui achieved 65.8 million transaction executions per day. This is the highest number in blockchain history. However, a single system can only integrate a limited number of cores, so the throughput gains from this strategy are limited. Moreover, machines that can accommodate large numbers of cores are expensive to purchase and use. The groundbreaking solution provided by Pilotfish allows a single validator to use multiple servers simultaneously. This new technology not only enables elastic capacity, but also allows smaller, cheaper servers to be used as needed through an auto-scaling architecture similar to centralized solutions like AWS or Heroku.

Almost all other blockchains that aim to scale use some form of batch processing, where transactions are grouped and added at once. The downside to this method is significant additional latency. Instead of transactions moving directly to the end, they are delayed as they wait for other transactions to be batched and executed. Unlike these other scaling solutions, Pilotfish achieves potentially linear throughput scaling without increasing latency. In the simplest terms, horizontal scaling enabled by Pilotfish allows validators to adjust throughput and resource usage to their current use case by simply using the required number of machines.

“Until now, the only way to increase the capacity of validators was to upgrade to more powerful machines,” said Alberto Sonnino, a researcher at Mysten Labs. Pilotfish removes these constraints, enabling horizontal scaling by simply adding servers without execution delays. With Pilotfish, it’s no longer a question of whether that level of throughput is possible, but only how many servers you need.”

Pilotfish achieves incredible scalability by allowing a single validator to execute transactions on multiple systems. Specifically, Pilotfish divides the verification function into three main roles: (1) a primary role that processes sequencing transactions using high-throughput consensus, (2) SequencingWorkers (SW), which store transactions and forward them for execution; (3) ExecutionWorkers (EW) that save the blockchain state and execute transactions received from SW. To increase system capacity more precisely, different components can be located in several different systems.

Experimental results show that Pilotfish has already achieved remarkable results in terms of linear scaling, low latency, and support for diverse workloads. In our tests, Pilotfish reduced network response times and kept latency below 20ms. Future plans include improvements such as multiple SequencingWorkers, shard replication, and ultra-fast networking.