Load Sharing in Distributed Systems

Wang, Yung-Terng; Morris, Robert

doi:10.1109/tc.1985.1676564

Cited by 332 publications

(3 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper, we consider the load balancing problem of assigning a stream of requests to a set of replicas of a service. There is no magic bullet, as the "best" choice of load balancing strategy depends on characteristics of the workload and system, as well as the desired outcome [10]. Arguably, the most commonly used strategies are round-robin (RR), random and join-shortest-queue (JSQ) (also known as Least-Connected/Least-Request) and their weighted counterparts [11,12], available in many modern software tools such as the proxies Envoy 3 and Nginx 4 .…”

Section: Load Balancingmentioning

confidence: 99%

Automatic Differentiation over Fluid Models for Holistic Load Balancing

Heimerson

Ruuskanen

Eker

2022

2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems Companion (ACSOS-C)

View full text Add to dashboard Cite

Microservice applications consist of a set of smaller services interacting in a graph structure to deliver the full application. Jobs will traverse this graph in different paths, both depending on the type of job, but also depending on the current load of different service replicas. Different paths will incur different scenario-specific costs, dependent on, e.g., deployment and the underlying cloud system. In this paper, we demonstrate how automatic differentiation over data-driven fluid models can be used to optimize a running microservice application, by designing a load balancer that minimizes some holistic cost function under response time constraints. First, a fluid model describing the load in each service is learned through parsing tracing data from the application. We introduce a cost function based on performance metrics such as mean queue length and response time percentiles, all retrieved from the fluid model. By using automatic differentiation on this cost function, we can find the gradient of the cost with respect to the load balancing parameters. This enables us to update these parameters, using e.g. gradient descent, in a manner that steers the application towards a more optimal setting. In an experimental evaluation on a small microservice application running on Ericsson Research Datacenter, it is shown that the method can quickly step towards optimal values while supporting complicated cost functions such as solutions to a system of ordinary differential equations.

show abstract

Section: Load Balancingmentioning

confidence: 99%

Automatic Differentiation over Fluid Models for Holistic Load Balancing

Heimerson

Ruuskanen

Eker

2022

2022 IEEE International Conference on Autonomic Computing and Self-Organizing Systems Companion (ACSOS-C)

View full text Add to dashboard Cite

show abstract

“…Unlike the static algorithms, the workload in dynamic algorithm is distributed among the servers at runtime. The balancer assigns a new user request to the servers based on the new information collected [4]. Dynamic algorithms allocate user requests dynamically when one of the servers becomes under loaded.…”

Section: Load Balancing Algorithmmentioning

confidence: 99%

An Adaptive Load Balancing Algorithm Based on Discrete Uniform Distribution

Shi

2010

AMR

View full text Add to dashboard Cite

Centralized cluster system is widely used by many web sites to increase service availability and balancing workload among multiple servers. In order to reduce the possibility which balancer becomes performance bottleneck in clusters, a dynamic adaptive weighted load-balanced algorithm is presented in this paper. The objective of the algorithm is that every working server sends a corresponding ask signals at different rates to balancer based on the discrete uniform distribution principle and the balancer assigning user requests according to an ask queue which consists of these signals and achieves the dynamic negative feedback on workload. Experiment results show that the algorithm is better than the static algorithm, as compared with the dynamic algorithm, it can also reduce the burden on load balancer.

show abstract

“…[5] compared two static workload allocation policies with six dynamic policies for several systems under a variety of loads. The evaluation of the different workload allocation policies was performed using 1 Heterogeneous in the context of this paper refers to differences in computational capability rather than differences in CPU architecture or operating system 2 For Quality-of-Service (QoS) demanding jobs, the mean miss rate is the average number of jobs that fail to meet the required QoS level. The profits under a Service Level Agreement (SLA) are maximised if all the jobs meet their QoS demands.…”

Section: Introductionmentioning

confidence: 99%

Traffic Offloading and Load Balancing to Enable Cloud Computing Connectivity

Shenfield

Fleming

2013

Int. J. Com. Net. Tech.

View full text Add to dashboard Cite

The distribution of computational tasks across a diverse set of geographically distributed heterogeneous resources is a critical issue in the realisation of true computational grids. Conventionally, workload allocation algorithms are divided into static and dynamic approaches. Whilst dynamic approaches frequently outperform static schemes, they usually require the collection and processing of detailed system information at frequent intervals -a task that can be both time consuming and unreliable in the realworld. This paper introduces a novel workload allocation algorithm for optimally distributing the workload produced by the arrival of batches of jobs. Results show that, for the arrival of batches of jobs, this workload allocation algorithm outperforms other commonly used algorithms in the static case. A hybrid scheduling approach (using this workload allocation algorithm), where information about the speed of computational resources is inferred from previously completed jobs, is then introduced and the efficiency of this approach demonstrated using a real world computational grid. These results are compared to the same workload allocation algorithm used in the static case and it can be seen that this hybrid approach comprehensively outperforms the static approach.

show abstract

Load Sharing in Distributed Systems

Cited by 332 publications

References 22 publications

Automatic Differentiation over Fluid Models for Holistic Load Balancing

Automatic Differentiation over Fluid Models for Holistic Load Balancing

An Adaptive Load Balancing Algorithm Based on Discrete Uniform Distribution

Traffic Offloading and Load Balancing to Enable Cloud Computing Connectivity

Contact Info

Product

Resources

About