Improving the Reactivity of Pure Operation-Based CRDTs

Research output: Chapter in Book/Report/Conference proceedingMeeting abstract (Book)

Abstract

Modern distributed applications increasingly replicate data to guarantee both high availability of the system and an optimal user experience. Conflict-Free Replicated Data Types (CRDTs) are a family of data types specially designed for highly available systems which guarantee some form of eventual consistency. In this paper, we show that the reliance on reliable causal broadcasting (RCB) middleware by existing CRDT frameworks may lead to less reactive CRDTs which in turn hampers user experience. We propose a solution that improves the reactivity of CRDTs built on an RCB middleware by reifying buffered operations. We apply our strategy to the pure operation-based CRDT framework, allowing for reactive pure operation-based CRDTs. We describe our approach as a formal extension to the framework and implement it in Flec, an extensible and open CRDT framework. The extension is then used to create new Add-Wins and Remove-Wins sets which exhibit higher reactivity.
Original languageEnglish
Title of host publicationPaPoC '21: Proceedings of the 8th Workshop on Principles and Practice of Consistency for Distributed Data
PublisherACM
Pages1-6
Number of pages6
ISBN (Electronic)9781450383387
ISBN (Print)978-1-4503-8338-7
DOIs
Publication statusPublished - Apr 2021
Event8th Workshop on Principles and Practice of Consistency for Distributed Data -
Duration: 26 Apr 2021 → …
https://papoc-workshop.github.io/2021/

Publication series

NameProceedings of the 8th Workshop on Principles and Practice of Consistency for Distributed Data, PaPoC 2021

Workshop

Workshop8th Workshop on Principles and Practice of Consistency for Distributed Data
Abbreviated titlePaPoC '21
Period26/04/21 → …
Internet address

Keywords

  • Garbage collection
  • Consistency
  • Distributed architectures
  • Synchronization

Fingerprint

Dive into the research topics of 'Improving the Reactivity of Pure Operation-Based CRDTs'. Together they form a unique fingerprint.

Cite this