The Performance of Paxos and Fast Paxos

TL;DR

This paper compares the performance of Paxos and Fast Paxos consensus algorithms using Treplica, revealing that Paxos performs better with fewer replicas and that collisions do not explain this difference.

Contribution

It provides an empirical performance comparison of Paxos and Fast Paxos in various failure scenarios using a modular replication toolkit.

Findings

Paxos outperforms Fast Paxos with small replica counts

Collisions are not responsible for performance differences

Performance varies between algorithms depending on failure conditions

Abstract

Paxos and Fast Paxos are optimal consensus algorithms that are simple and elegant, while suitable for efficient implementation. In this paper, we compare the performance of both algorithms in failure-free and failure-prone runs using Treplica, a general replication toolkit that implements these algorithms in a modular and efficient manner. We have found that Paxos outperforms Fast Paxos for small number of replicas and that collisions are not the cause of this performance difference.