The Power of d Choices for Redundancy

An increasingly prevalent technique for improving response time in queueing systems is the use of redundancy. In a system with redundant requests, each job that arrives to the system is copied and dispatched to multiple servers. As soon as the first copy completes service, the job is considered complete, and all remaining copies are deleted. A great deal of empirical work has demonstrated that redundancy can significantly reduce response time in systems ranging from Google's BigTable service to kidney transplant waitlists. We propose a theoretical model of redundancy, the Redundancy-d system, in which each job sends redundant copies to d servers chosen uniformly at random. We derive the first exact expressions for mean response time in Redundancy-d systems with any finite number of servers. We also find asymptotically exact expressions for the distribution of response time as the number of servers approaches infinity.

[1]  Albert G. Greenberg,et al.  Reining in the Outliers in Map-Reduce Clusters using Mantri , 2010, OSDI.

[2]  Alan Scheller-Wolf,et al.  A Better Model for Job Redundancy: Decoupling Server Slowdown and Job Size , 2016, IEEE/ACM Transactions on Networking.

[3]  Mor Harchol-Balter,et al.  Reducing Latency via Redundant Requests: Exact Analysis , 2015, SIGMETRICS 2015.

[4]  Robert M Merion,et al.  Prevalence and Outcomes of Multiple‐Listing for Cadaveric Kidney and Liver Transplantation , 2004, American journal of transplantation : official journal of the American Society of Transplantation and the American Society of Transplant Surgeons.

[5]  Randy H. Katz,et al.  Improving MapReduce Performance in Heterogeneous Environments , 2008, OSDI.

[6]  Michael Mitzenmacher,et al.  The Power of Two Choices in Randomized Load Balancing , 2001, IEEE Trans. Parallel Distributed Syst..

[7]  Emina Soljanin,et al.  Coding for fast content download , 2012, 2012 50th Annual Allerton Conference on Communication, Control, and Computing (Allerton).

[8]  Yi Lu,et al.  Asymptotic independence of queues under randomized load balancing , 2012, Queueing Syst. Theory Appl..

[9]  R. L. Dobrushin,et al.  Queueing system with selection of the shortest of two queues: an assymptotic approach , 1996 .

[10]  Mor Harchol-Balter,et al.  Analyzing Response Time in the Redundancy-d System , 2015 .

[11]  Sanjay Ghemawat,et al.  MapReduce: Simplified Data Processing on Large Clusters , 2004, OSDI.

[12]  Sridhar R. Tayur,et al.  OrganJet: Overcoming Geographical Disparities in Access to Deceased Donor Kidneys in the United States , 2017, Manag. Sci..

[13]  Kannan Ramchandran,et al.  The MDS Queue: Analysing Latency Performance of Codes and Redundant Requests , 2012 .

[14]  Ger Koole,et al.  Resource allocation in grid computing , 2008, J. Sched..

[15]  R. Srikant,et al.  The power of slightly more than one sample in randomized load balancing , 2015, 2015 IEEE Conference on Computer Communications (INFOCOM).

[16]  Luiz André Barroso,et al.  The tail at scale , 2013, CACM.

[17]  Mohammad Sanaei Ardekani,et al.  Multiple listing in kidney transplantation. , 2010, American journal of kidney diseases : the official journal of the National Kidney Foundation.

[18]  Brian D. Noble,et al.  Small is better: avoiding latency traps in virtualized data centers , 2013, SoCC.

[19]  Scott Shenker,et al.  Usenix Association 10th Usenix Symposium on Networked Systems Design and Implementation (nsdi '13) 185 Effective Straggler Mitigation: Attack of the Clones , 2022 .

[20]  Emina Soljanin,et al.  On the Delay-Storage Trade-Off in Content Download from Coded Distributed Storage Systems , 2013, IEEE Journal on Selected Areas in Communications.