vix.ing · top · new · best · stats · spec

XPC: Fast and Reliable Synchronous Transmission Protocols for 2-Phase Commit and 3-Phase Commit

2019/10/22 by Alberto Spina, Spina, Alberto, Michael Breza +6
Computer Science · #Distributed #Distributed systems and fault tolerance #FOS: Computer and information sciences #Interconnection Networks and Systems #Networking and Internet Architecture (cs.NI) #Parallel #Real-Time Systems Scheduling #and Cluster Computing (cs.DC) #cs.DC #cs.NI

paper · pdf · doi:10.48550/arxiv.1910.09941

13 Pages

arxiv created 2019/10/22 · openalex publication_date 2019/10/22 · arxiv updated 2019/10/23 · openalex created_date 2019/11/01 · openalex updated_date 2026/07/28

Abstract

One of the major challenges for the engineering of wireless sensing systems is to improve the software abstractions and frameworks that are available to programmers while ensuring system reliability and efficiency. The distributed systems community have developed a rich set of such abstractions for building dependable distributed systems connected using wired networks, however after 20 years research many of these elude wireless sensor systems. In this paper we present X Process Commit (XPC) an atomic commit protocol framework that utilizes Synchronous Transmission (ST). We also introduce Hybrid, a technique that allows us to exploit the advantages of the Glossy and Chaos Synchronous Transmission primitives to get lower latency and higher reliability than either. Using XPC and Hybrid we demonstrate how to build protocols for the classical 2-phase and 3-phase commit abstractions and evaluate these demonstrating significantly improved performance and reliability than the use of Glossy or Chaos individually as dissemination primitives. We address how we overcame the timing challenges of bringing Glossy and Chaos together to form Hybrid and through extensive experimentation demonstrate that it is robust to in-network radio interference caused by multiple sources. We are first to present testbed results that show that Hybrid can provide almost 100% reliability in a network of nodes suffering from various levels of radio interference.

Related