CaltechAUTHORS
  A Caltech Library Service

GRASS: Trimming Stragglers in Approximation Analytics

Ananthanarayanan, Ganesh and Hung, Michael Chien-Chun and Ren, Xiaoqi and Stoica, Ion and Wierman, Adam and Yu, Minlan (2014) GRASS: Trimming Stragglers in Approximation Analytics. In: Proceedings of the 11th USENIX Symposium on Networked Systems Design and Implementation (NSDI ’14). USENIX Association , Berkeley, CA, pp. 289-302. ISBN 978-1-931971-09-6. http://resolver.caltech.edu/CaltechAUTHORS:20160419-161606584

[img] PDF (Paper) - Published Version
See Usage Policy.

797Kb
[img] PDF (Slides) - Published Version
See Usage Policy.

1843Kb
[img] Video (MPEG) (Presentation Video) - Presentation
See Usage Policy.

123Mb

Use this Persistent URL to link to this item: http://resolver.caltech.edu/CaltechAUTHORS:20160419-161606584

Abstract

In big data analytics, timely results, even if based on only part of the data, are often good enough. For this reason, approximation jobs, which have deadline or error bounds and require only a subset of their tasks to complete, are projected to dominate big data workloads. Straggler tasks are an important hurdle when designing approximate data analytic frameworks, and the widely adopted approach to deal with them is speculative execution. In this paper, we present GRASS, which carefully uses speculation to mitigate the impact of stragglers in approximation jobs. GRASS’s design is based on first principles analysis of the impact of speculation. GRASS delicately balances immediacy of improving the approximation goal with the long term implications of using extra resources for speculation. Evaluations with production workloads from Facebook and Microsoft Bing in an EC2 cluster of 200 nodes shows that GRASS increases accuracy of deadline-bound jobs by 47% and speeds up error-bound jobs by 38%. GRASS’s design also speeds up exact computations (zero error-bound), making it a unified solution for straggler mitigation.


Item Type:Book Section
Related URLs:
URLURL TypeDescription
https://www.usenix.org/conference/nsdi14/technical-sessions/presentation/ananthanarayananOrganizationPaper
Additional Information:© 2014 Usenix Association. We thank our shepherd Nina Taft and the anonymous reviewers for their suggestions to improve this work. We also thank Rohan Gandhi for his feedback on our early drafts. This research was partially funded by research grant NSF CNS-1319820, NSF CISE Expeditions award CCF-1139158, the DARPA XData Award FA8750-12-2-0331, and gifts from Qualcomm, Amazon Web Services, Google, SAP, Blue Goji, Cisco, Clearstory Data, Cloudera, Ericsson, Facebook, General Electric, Hortonworks, Huawei, Intel, Microsoft, NetApp, Oracle, Quanta, Samsung, Splunk, VMware and Yahoo!.
Funders:
Funding AgencyGrant Number
NSFCNS-1319820
NSFCCF-1139158
Defense Advanced Research Projects Agency (DARPA)UNSPECIFIED
QualcommUNSPECIFIED
Amazon Web ServicesUNSPECIFIED
GoogleUNSPECIFIED
SAPUNSPECIFIED
Blue GojiUNSPECIFIED
CiscoUNSPECIFIED
Clearstory DataUNSPECIFIED
ClouderaUNSPECIFIED
EricssonUNSPECIFIED
FacebookUNSPECIFIED
General ElectricUNSPECIFIED
HortonworksUNSPECIFIED
HuaweiUNSPECIFIED
IntelUNSPECIFIED
MicrosoftUNSPECIFIED
NetAppUNSPECIFIED
OracleUNSPECIFIED
QuantaUNSPECIFIED
SamsungUNSPECIFIED
SplunkUNSPECIFIED
VMwareUNSPECIFIED
Yahoo!UNSPECIFIED
Air Force Office of Scientific Research (AFOSR)FA8750-12-2-0331
Record Number:CaltechAUTHORS:20160419-161606584
Persistent URL:http://resolver.caltech.edu/CaltechAUTHORS:20160419-161606584
Usage Policy:No commercial reproduction, distribution, display or performance rights in this work are provided.
ID Code:66286
Collection:CaltechAUTHORS
Deposited By: Kristin Buxton
Deposited On:19 Apr 2016 23:28
Last Modified:19 Apr 2016 23:28

Repository Staff Only: item control page