Abstract
In multi-server systems, a classical job assignment algorithm works as follows: at the arrival of each job, pick d servers independently and uniformly at random and send the job to the least loaded server among the d servers. This model is known as the power-of-d choices algorithm. In this paper, we analyze a variant of this algorithm, where d servers are sampled through d independent non-backtracking random walks on a k -regular graph. The random walkers are periodically reset to independent uniform random positions. Under some assumptions on the underlying graph, we show that the system dynamics under this new algorithm converges to the solution of a deterministic ordinary differential equation (ODE), which is the same ODE as the classical power-of-d choices. We also show that the new algorithm stablizes the system, and the stationary distribution of the system converges to the stationary solution of the ODE. The new scheme can be considered as a derandomized version of power-of- d choices as it reduces the use of randomness while maintaining the performance of power-of- d choices.
- Reza Aghajani and Kavita Ramanan. 2017. The hydrodynamic limit of a randomized load balancing network. arXiv preprint arXiv:1707.02005 (2017).Google Scholar
- Noga Alon, Itai Benjamini, Eyal Lubetzky, and Sasha Sodin. 2007. Non-backtracking random walks mix faster. Communications in Contemporary Mathematics , Vol. 9, 04 (2007), 585--603.Google Scholar
Cross Ref
- Noga Alon and Eyal Lubetzky. 2009. Poisson approximation for non-backtracking random walks. Israel Journal of Mathematics , Vol. 174, 1 (2009), 227--252.Google Scholar
Cross Ref
- Patrick Billingsley. 2013. Convergence of probability measures .John Wiley & Sons.Google Scholar
- Aleksandr Alekseevich Borovkov. 1998. Ergodicity and stability of stochastic processes. (1998).Google Scholar
- Amarjit Budhiraja, Debankur Mukherjee, and Ruoyu Wu. 2017. Supermarket model on graphs. arXiv preprint arXiv:1712.07607 (2017).Google Scholar
- Fan Chung and Linyuan Lu. 2006. Concentration inequalities and martingale inequalities: a survey. Internet Mathematics , Vol. 3, 1 (2006), 79--127.Google Scholar
Cross Ref
- R.W.R. Darling and James R. Norris. 2008. Differential equation approximations for Markov chains. Probability surveys , Vol. 5 (2008), 37--79.Google Scholar
- Moez Draief and Laurent Massouli. 2010. Epidemics and rumours in complex networks .Cambridge University Press. Google Scholar
Digital Library
- Stewart N. Ethier and Thomas G. Kurtz. 2009. Markov processes: characterization and convergence. Vol. 282. John Wiley & Sons.Google Scholar
- G. Foschini and Jack Salz. 1978. A basic dynamic routing problem and diffusion. IEEE Transactions on Communications , Vol. 26, 3 (1978), 320--327.Google Scholar
Cross Ref
- Ayalvadi Ganesh, Sarah Lilienthal, D Manjunath, Alexandre Proutiere, and Florian Simatos. 2010. Load balancing via random local search in closed and open systems. In ACM SIGMETRICS Performance Evaluation Review, Vol. 38. ACM, 287--298. Google Scholar
Digital Library
- Nicolas Gast. 2015. The Power of Two Choices on Graphs: the Pair-Approximation is Accurate? ACM SIGMETRICS Performance Evaluation Review , Vol. 43, 2 (2015), 69--71. Google Scholar
Digital Library
- Nicolas Gast. 2017. Expected values estimated via mean-field approximation are 1/N-accurate. Proceedings of the ACM on Measurement and Analysis of Computing Systems , Vol. 1, 1 (2017), 17. Google Scholar
Digital Library
- Michel Habib, Colin McDiarmid, Jorge Ramirez-Alfonsin, and Bruce Reed. 2013. Probabilistic methods for algorithmic discrete mathematics. Vol. 16. Springer Science & Business Media.Google Scholar
- Shlomo Hoory, Nathan Linial, and Avi Wigderson. 2006. Expander graphs and their applications. Bull. Amer. Math. Soc. , Vol. 43, 4 (2006), 439--561.Google Scholar
Cross Ref
- Joohwan Kim and R Srikant. 2013. Real-time peer-to-peer streaming over multiple random hamiltonian cycles. IEEE Transactions on Information Theory , Vol. 59, 9 (2013), 5763--5778. Google Scholar
Digital Library
- Alexander Lubotzky, Ralph Phillips, and Peter Sarnak. 1988. Ramanujan graphs. Combinatorica , Vol. 8, 3 (1988), 261--277.Google Scholar
Cross Ref
- MJ Luczak and JR Norris. 2013. Averaging over fast variables in the fluid limit for Markov chains: application to the supermarket model with memory. The Annals of Applied Probability , Vol. 23, 3 (2013), 957--986.Google Scholar
Cross Ref
- Malwina J Luczak and Colin McDiarmid. 2006. On the maximum queue length in the supermarket model. The Annals of Probability , Vol. 34, 2 (2006), 493--527.Google Scholar
Cross Ref
- Malwina J Luczak and James Norris. 2005. Strong approximation for the supermarket model. The Annals of Applied Probability , Vol. 15, 3 (2005), 2038--2061.Google Scholar
Cross Ref
- Michael Mitzenmacher. 1996. Load balancing and density dependent jump Markov processes. In IEEE Symposium on Foundations of Computer Science (FOCS). IEEE, 213. Google Scholar
Digital Library
- Michael Mitzenmacher. 2001. The power of two choices in randomized load balancing. IEEE Transactions on Parallel and Distributed Systems , Vol. 12, 10 (2001), 1094--1104. Google Scholar
Digital Library
- Michael Mitzenmacher, Balaji Prabhakar, and Devavrat Shah. 2002. Load balancing with memory. In IEEE Symposium on Foundations of Computer Science (FOCS). IEEE, 799. Google Scholar
Digital Library
- Debankur Mukherjee, Sem C Borst, and Johan SH van Leeuwaarden. 2018. Asymptotically optimal load balancing topologies. Proceedings of the ACM on Measurement and Analysis of Computing Systems , Vol. 2, 1 (2018), 14. Google Scholar
Digital Library
- Debankur Mukherjee, Sem C Borst, Johan SH van Leeuwaarden, and Philip A Whiting. 2016. Universality of Power-of-$ d $ Load Balancing in Many-Server Systems. arXiv preprint arXiv:1612.00723 (2016).Google Scholar
- Alain-Sol Sznitman. 1991. Topics in propagation of chaos. In Ecole d'été de probabilités de Saint-Flour XIX--1989. Springer, 165--251.Google Scholar
- Dengwang Tang and Vijay G. Subramanian. 2018. Balanced Allocation with Random Walk Based Sampling. arXiv preprint arXiv:1810.02722 (2018).Google Scholar
- Nikita Dmitrievna Vvedenskaya, Roland L'vovich Dobrushin, and Fridrikh Izrailevich Karpelevich. 1996. Queueing system with selection of the shortest of two queues: An asymptotic approach. Problemy Peredachi Informatsii , Vol. 32, 1 (1996), 20--34.Google Scholar
- Nicholas C Wormald. 1995. Differential equations for random processes and random graphs. The annals of applied probability , Vol. 5, 4 (1995), 1217--1235.Google Scholar
- Lei Ying. 2016. On the approximation error of mean-field models. In ACM SIGMETRICS Performance Evaluation Review , Vol. 44. ACM, 285--297. Google Scholar
Digital Library
- Lei Ying, R. Srikant, and Xiaohan Kang. 2015. The power of slightly more than one sample in randomized load balancing. In Computer Communications (INFOCOM), 2015 IEEE Conference on. IEEE, 1131--1139.Google Scholar
Cross Ref
- Se-Young Yun and Alexandre Proutiere. 2015. Distributed proportional fair load balancing in heterogenous systems. In ACM SIGMETRICS Performance Evaluation Review, Vol. 43. ACM, 17--30. Google Scholar
Digital Library
Index Terms
Random Walk Based Sampling for Load Balancing in Multi-Server Systems
Recommendations
Random Walk Based Sampling for Load Balancing in Multi-Server Systems
SIGMETRICS '19: Abstracts of the 2019 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer SystemsIn multi-server systems, a classical job assignment algorithm works as follows: at the arrival of each job, pick d servers independently and uniformly at random and send the job to the least loaded server among the d servers. This model is known as the ...
Random Walk Based Sampling for Load Balancing in Multi-Server Systems
In multi-server systems, a classical job assignment algorithm works as follows: at the arrival of each job, pick d servers independently and uniformly at random and send the job to the least loaded server among the d servers. This model is known as the ...
Study of multi server retrial queueing system under vacation policies by direct truncation method
QTNA '11: Proceedings of the 6th International Conference on Queueing Theory and Network ApplicationsConsider a Multi server Retrial queueing system with vacation policies in which arrival rate follows a Poisson distribution with parameter λ and service time follows an exponential distribution with parameter μ. Let c be the number of servers in the ...






Comments