Mean field approximations to a queueing system with threshold-based workload control scheme

In this paper, motivated by considerations of server utilization and energy consumptions in cloud computing, we investigate a homogeneous queueing system with a threshold-based workload control scheme. In this system, a virtual machine will be turned off when there are no tasks in its buffer upon the completion of a service by the machine, and turned on when the number of tasks in its buffer reaches a pre-set threshold value. Due to complexity of this system, we propose approximations to system performance measures by mean field limits. An iterative algorithm is suggested for the solution to the mean field limit equations. In addition, numerical and simulation results are presented to justify the proposed approximation method and to provide a numerical analysis on the impact of the system performances by system parameters.

[1]  Balamurugan Balusamy,et al.  Threshold-Based Workload Control for an Under-Utilized Virtual Machine in Cloud Computing , 2016 .

[2]  John C. S. Lui,et al.  The mean-field computation in a supermarket model with server multiple vacations , 2013, Discrete Event Dynamic Systems.

[3]  Yu. M. Suhov,et al.  Fast Jackson networks , 1999 .

[4]  Tonghoon Suk,et al.  Randomized longest-queue-first scheduling for large-scale buffered systems , 2013, Advances in Applied Probability.

[5]  F. Richard Yu,et al.  Cloud radio access networks (C-RAN) in mobile cloud computing systems , 2014, 2014 IEEE Conference on Computer Communications Workshops (INFOCOM WKSHPS).

[6]  Sem C. Borst,et al.  Optimal Service Elasticity in Large-Scale Distributed Systems , 2017, SIGMETRICS.

[7]  Ravi Mazumdar,et al.  Choosing among heterogeneous server clouds , 2016, Queueing Systems.

[8]  Quanlin Li,et al.  Mean-Field Analysis for Heterogeneous Work Stealing Models , 2015 .

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

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

[11]  Arpan Mukhopadhyay,et al.  Mean Field Interactions in Heterogeneous Networks , 2016 .

[12]  Dongyu Qiu,et al.  Modeling of the resource allocation in cloud computing centers , 2015, Comput. Networks.

[14]  H. McKean,et al.  A CLASS OF MARKOV PROCESSES ASSOCIATED WITH NONLINEAR PARABOLIC EQUATIONS , 1966, Proceedings of the National Academy of Sciences of the United States of America.

[15]  Yiqiang Q. Zhao,et al.  Balancing Queues by Mean Field Interaction , 2005, Queueing Syst. Theory Appl..

[16]  R. Srikant,et al.  The Power of Slightly More than One Sample in Randomized Load Balancing , 2017, Math. Oper. Res..