Electronic Communications in Probability

Load balancing under $d$-thinning

Ohad Noy Feldheim and Jiange Li

Full-text: Open access

Abstract

In the classical balls-and-bins model, $m$ balls are allocated into $n$ bins one by one uniformly at random. In this note, we consider the $d$-thinning variant of this model, in which the process is regulated in an on-line fashion as follows. For each ball, after a random bin has been selected, an overseer may decide, based on all previous history, whether to accept this bin or not. However, one of every $d$ consecutive suggested bins must be accepted. The maximum load of this setting is the number of balls in the most loaded bin. We show that after $\Theta (n)$ balls have been allocated, the least maximum load achievable with high probability is $(d+o(1))\sqrt [d]{\frac{d\log n} {\log \log n}}$. This should be compared with the related $d$-choice setting, in which the optimal maximum load achievable with high probability is $\frac{\log \log n} {\log d}+O(1)$.

Article information

Source
Electron. Commun. Probab., Volume 25 (2020), paper no. 1, 13 pp.

Dates
Received: 27 August 2019
Accepted: 9 December 2019
First available in Project Euclid: 3 January 2020

Permanent link to this document
https://projecteuclid.org/euclid.ecp/1578020668

Digital Object Identifier
doi:10.1214/19-ECP282

Zentralblatt MATH identifier
07149364

Subjects
Primary: 60J10: Markov chains (discrete-time Markov processes on discrete state spaces) 68Q87: Probability in computer science (algorithm analysis, random structures, phase transitions, etc.) [See also 68W20, 68W40]

Keywords
balls and bins load balancing two-choice thinning

Rights
Creative Commons Attribution 4.0 International License.

Citation

Feldheim, Ohad Noy; Li, Jiange. Load balancing under $d$-thinning. Electron. Commun. Probab. 25 (2020), paper no. 1, 13 pp. doi:10.1214/19-ECP282. https://projecteuclid.org/euclid.ecp/1578020668


Export citation

References

  • [1] Adler, M., Chakrabarti, S., Mitzenmacher, M. and Rasmussen, L.: Parallel randomized load balancing, Random Struct. Algor., 13(2), (1998), 159–188.
  • [2] Azar, Y., Broder, A., Karlin, A. and Upfal, E.: Balanced allocations, SIAM Journal of Computing, 29(1), 1999, 180–200.
  • [3] Dwivedi, R., Feldheim, O. N., Gurel-Gurevich, O. and Ramdas, A.: The power of online thinning in reducing discrepancy, Probab. Theory Related Fields., 174(1-2), 2019, 103–131.
  • [4] Feldheim, O. N. and Gurel-Gurevich, O.: The power of thinning in balanced allocation, Preprint, arXiv:1807.01132.
  • [5] Karp, R. M., Luby, M. and Meyer auf der Heide, F.: Efficient PRAM simulation on a distributed memory machine, Algorithmica, 16, 1996, 517–542.
  • [6] Korst, J.: Random duplicated assignment: An alternative to striping in video servers, In Proceedings of the 5th ACM International Multimedia Conference, Seattle, WA, ACM Press, New York, 1997, 219–226.
  • [7] Mitzenmacher, M.: The power of two choices in randomized load balancing, PhD thesis. University of California, Berkeley, 1996.
  • [8] Mitzenmacher, M. and Upfal, E.: Probability and computing: Randomized algorithms and probabilistic analysis, Cambridge University Press, 2005.
  • [9] Peres, Y., Talwar, K. and Wieder, U.: Graphical balanced allocations and the $(1+\beta )$-choice process, Random Struct. Algor., 47(4), 2015, 760–775.
  • [10] Raab, M. and Steger, A.: Balls into bins–a simple and tight analysis, In Proceedings of the 2nd International Workshop on Randomization and Approximation Techniques in Computer Science, Barcelona, Spain, 1998.
  • [11] Sanders, P., Egner, S. and Korst, J.: Fast concurrent access to parallel disks, Algorithmica, 35, 2003, 21–55.
  • [12] Stemann, V.: Parallel balanced allocations, In Proceedings of the 8th Annual ACM Symposium on Parallel Algorithms and Architectures, Padua, Italy, ACM Press, New York, 1996, 261–269.