Users of modern distributed stream processing systems have to choose between non-deterministic computations and high latency due to a need in excessive buffering. We introduce a speculative model based on MapReduce-complete set of operations that allows us to achieve determinism and low-latency. Experiments show that our prototype can outperform existing solutions due to low overhead of optimistic synchronization.

Original languageEnglish
Title of host publicationAdvances in Databases and Information Systems - 22nd European Conference, ADBIS 2018, Proceedings
EditorsAndras Benczur, Tomas Horvath, Bernhard Thalheim
PublisherSpringer Nature
Pages233-246
Number of pages14
ISBN (Print)9783319983974
DOIs
StatePublished - 1 Jan 2018
Event22nd East-European Conference on Advances in Databases and Information Systems, ADBIS 2018 - Budapest, Hungary
Duration: 2 Sep 20185 Sep 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11019 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference22nd East-European Conference on Advances in Databases and Information Systems, ADBIS 2018
Country/TerritoryHungary
CityBudapest
Period2/09/185/09/18

    Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

    Research areas

  • Data streams, Distributed processing, Drifting state

ID: 36626901