Balanced fair resource sharing in computer clusters

Bonald, Thomas; Comte, Céline

doi:10.1016/j.peva.2017.08.006

Cited by 44 publications

(58 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…single server. On the other hand, [9,15,4] focus on the problem of resource allocation, assuming a static load balancing that assigns incoming jobs to classes at random, independently of the system state. The class of a job in the system identifies the set of servers that can be pooled to process it in parallel.…”

Section: Serversmentioning

confidence: 99%

“…Depending on the system state, the load balancer matches the job with a compatible class that subsequently determines its assigned servers. The upper part of our graph, which puts constraints on load balancing, corresponds to the bipartite graph of [1]; the lower part, which restricts the resource allocation, corresponds to the bipartite graph of [9,15,4].…”

Section: Serversmentioning

confidence: 99%

“…The rationale behind this algorithm is to use the release order of tokens as an information on the relative load of their servers: a token that has been available for a long time without being seized is likely to identify a server set that is less loaded than others. As we will see, our algorithm mirrors the first-come, first-served (FCFS) service discipline proposed in [4] to implement balanced fairness, which was defined in [7] as the most efficient insensitive resource allocation.…”

Section: Serversmentioning

confidence: 99%

“…We first recall the model considered in [9,15,4] to study the problem of resource allocation in server pools. This model will be extended in Section 3 to integrate dynamic load balancing.…”

Section: Resource Allocationmentioning

confidence: 99%

“…Some recent works have tackled this problem from the point of view of queueing theory [1,15,9,4]. Their common feature is the adoption of a bipartite graph that translates practical constraints such as data locality into compatibility relations between jobs and servers.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Dynamic Load Balancing with Tokens

Comte¹

2018

2018 IFIP Networking Conference (IFIP Networking) and Workshops

View full text Add to dashboard Cite

Efficiently exploiting the resources of data centers is a complex task that requires efficient and reliable load balancing and resource allocation algorithms. The former are in charge of assigning jobs to servers upon their arrival in the system, while the latter are responsible for sharing server resources between their assigned jobs. These algorithms should take account of various constraints, such as data locality, that restrict the feasible job assignments. In this paper, we propose a token-based mechanism that efficiently balances load between servers without requiring any knowledge on job arrival rates and server capacities. Assuming a balanced fair sharing of the server resources, we show that the resulting dynamic load balancing is insensitive to the job size distribution. Its performance is compared to that obtained under the best static load balancing and in an ideal system that would constantly optimize the resource utilization. * c IFIP, 2018. This is the author's version of the work. It is posted here by permission of IFIP for your personal use. Not for redistribution. The definitive version will be published in IFIP NETWORKING 2018 proceedings.

show abstract

Section: Serversmentioning

confidence: 99%

Section: Serversmentioning

confidence: 99%

Section: Serversmentioning

confidence: 99%

“…We first recall the model considered in [9,15,4] to study the problem of resource allocation in server pools. This model will be extended in Section 3 to integrate dynamic load balancing.…”

Section: Resource Allocationmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Dynamic Load Balancing with Tokens

Comte¹

2018

2018 IFIP Networking Conference (IFIP Networking) and Workshops

View full text Add to dashboard Cite

show abstract

A Survey of Stability Results for Redundancy Systems

Anton

Ayesta

Jonckheere

et al. 2021

Emergence, Complexity and Computation

View full text Add to dashboard Cite

Poly-symmetry in processor-sharing systems

Bonald

Comte²,

Shah³

et al. 2017

Queueing Syst

Self Cite

View full text Add to dashboard Cite

We consider a system of processor-sharing queues with state-dependent service rates. These are allocated according to balanced fairness within a polymatroid capacity set. Balanced fairness is known to be both insensitive and Pareto-efficient in such systems, which ensures that the performance metrics, when computable, will provide robust insights into the real performance of the system considered. We first show that these performance metrics can be evaluated with a complexity that is polynomial in the system size if the system is partitioned into a finite number of parts, so that queues are exchangeable within each part and asymmetric across different parts. This in turn allows us to derive stochastic bounds for a larger class of systems which satisfy less restrictive symmetry assumptions. These results are applied to practical examples of tree data networks, such as backhaul networks of Internet service providers, and computer clusters.

show abstract

Balanced fair resource sharing in computer clusters

Cited by 44 publications

References 31 publications

Dynamic Load Balancing with Tokens

Dynamic Load Balancing with Tokens

A Survey of Stability Results for Redundancy Systems

Poly-symmetry in processor-sharing systems

Contact Info

Product

Resources

About