New techniques to curtail the tail latency in stream processing systems

Guangxiang Du, Indranil Gupta

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

This paper presents a series of novel techniques for reducing the tail latency in stream processing systems like Apache Storm. Concretely, we present three mechanisms: (1) adaptive timeout coupled with selective replay to catch straggler tuples; (2) shared queues among different tasks of the same operator to reduce overall queueing delay; (3) latency feedback-based load balancing, intended to mitigate het-erogenous scenarios. We have implemented these techniques in Apache Storm, and present experimental results using sets of micro-benchmarks as well as two topologies from Yahoo! Inc. Our results show improvement in tail latency up to 72.9%.

Original languageEnglish (US)
Title of host publicationProceedings of the 4th Workshop on Distributed Cloud Computing, DCC 2016
PublisherAssociation for Computing Machinery
ISBN (Print)9781450342209
DOIs
StatePublished - Jul 25 2016
Event4th Annual ACM PODC Workshop on Distributed Cloud Computing, DCC 2016 - Chicago, United States
Duration: Jul 25 2016Jul 28 2016

Publication series

NameProceedings of the Annual ACM Symposium on Principles of Distributed Computing

Other

Other4th Annual ACM PODC Workshop on Distributed Cloud Computing, DCC 2016
Country/TerritoryUnited States
CityChicago
Period7/25/167/28/16

Keywords

  • Apache Storm
  • Stream Processing Systems
  • Tail Latency

ASJC Scopus subject areas

  • Software
  • Hardware and Architecture
  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'New techniques to curtail the tail latency in stream processing systems '. Together they form a unique fingerprint.

Cite this