English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

With the emergence of large-scale decentralized applications, a scalable andefficient Byzantine Fault Tolerant (BFT) protocol of hundreds of replicas isdesirable. Although the throughput of existing leader-based BFT protocols hasreached a high level of $10^5$ requests per second for a small scale ofreplicas, it drops significantly when the number of replicas increases, whichleads to a lack of practicality. This paper focuses on the scalability of BFTprotocols and identifies a major bottleneck to leader-based BFT protocols dueto the excessive workload of the leader at large scales. A new metric ofscaling factor is defined to capture whether a BFT protocol will get stuck whenit scales out, which can be used to measure the performance of efficiency andscalability of BFT protocols. We propose "Leopard", the first leader-based BFTprotocol that scales to multiple hundreds of replicas, and more importantly,preserves a high efficiency. We remove the bottleneck by introducing atechnique of achieving a constant scaling factor, which takes full advantage ofthe idle resource and adaptively balances the workload of the leader among allreplicas. We implement Leopard and evaluate its performance compared toHotStuff, the state-of-the-art BFT protocol. We run extensive experiments onthe two systems with up to 600 replicas. The results show that Leopard achievessignificant performance improvements both on throughput and scalability. Inparticular, the throughput of Leopard remains at a high level of $10^5$ whenthe system scales out to 600 replicas. It achieves a $5\times$ throughput overHotStuff when the scale is 300 (which is already the largest scale we can seethe progress of the latter in our experiments), and the gap becomes wider asthe number of replicas further increases.

Leopard: Towards High Throughput-Preserving BFT for Large-scale Systems