We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:


Current browse context:


Change to browse by:

References & Citations

DBLP - CS Bibliography


(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Networking and Internet Architecture

Title: Load Balancing in Heterogeneous Server Clusters: Insights From a Product-Form Queueing Model

Authors: Mark van der Boor (TU/e), Céline Comte (TU/e)
Abstract: Efficiently exploiting servers in data centers requires performance analysis methods that account not only for the stochastic nature of demand but also for server heterogeneity. Although several recent works proved optimality results for heterogeneity-aware variants of classical load-balancing algorithms in the many-server regime, we still lack a fundamental understanding of the impact of heterogeneity on performance in finite-size systems. In this paper, we consider a load-balancing algorithm that leads to a product-form queueing model and can therefore be analyzed exactly even when the number of servers is finite. We develop new analytical methods that exploit its product-form stationary distribution to understand the joint impact of the speeds and buffer lengths of servers on performance. These analytical results are supported and complemented by numerical evaluations that cover a large variety of scenarios.
Subjects: Networking and Internet Architecture (cs.NI); Probability (math.PR)
Journal reference: 2021 IEEE/ACM 29th International Symposium on Quality of Service (IWQOS), Jun 2021, Tokyo, Japan. pp.1-10
DOI: 10.1109/IWQOS52092.2021.9521355
Cite as: arXiv:2109.00868 [cs.NI]
  (or arXiv:2109.00868v1 [cs.NI] for this version)

Submission history

From: Celine Comte [view email]
[v1] Thu, 2 Sep 2021 12:30:43 GMT (45kb)

Link back to: arXiv, form interface, contact.