Symmetric Peer-to-Peer Applications

Abstract

In real-time networked applications, such as collaborative documents, users can interact remotelyand yet share the exact same experience as if they were using a single machine. In this work,we propose a middleware for symmetric peer-to-peer applications in which decentralized instancescan broadcast asynchronous events and yet conform to identical behavior. Peers are allowed tojoin and leave the network at any time. Application developers must adhere to a determinis-tic and stateless API. The middleware is responsible for synchronizing the events in a globalshared timeline across the network. Also, based on memory snapshots and deterministic simu-lation, a “time machine” can rollback conflicting peers to resynchronize their states. We showthat the middleware can handle applications with over 20 peers in a heterogeneous topologiesunder high churns. For instance, in a simulation scenario with a network latency of 50ms and 25 events per minute, the peers need to roll back and resynchronize only 3% of the time.