Load Balancing for Response Time

Journal of Parallel and Distributed Computing, 2011

This paper diverges from the traditional load balancing, and introduces a new principle called the onmachine load balance rule. The on-machine load balance rule leads to resource allocations that are better in tolerating uncertainties in the processing times of the tasks allocated to the resources when compared to other resource allocations that are derived using the conventional ''across-the-machines'' load balancing rule. The on-machine load balance rule calls for the resource allocation algorithms to allocate similarly sized tasks on a machine (in addition to optimizing some primary performance measures such as estimated makespan and average response time). The on-machine load balance rule is very different from the usual across-the-machines load balance rule that strives to balance load across resources so that all resources have similar finishing times. We give a mathematical justification for the on-machine load balance rule requiring only liberal assumptions about task processing times. Then we validate with extensive simulations that the resource allocations derived using on-machine load balance rule are indeed more tolerant of uncertain task processing times.

1996

We consider the following load balancing problem. Jobs arrive on-line and must be assigned to one of m machines thereby increasing the load on that machine by a certain weight. Jobs also depart on-line. The goal is to minimize the maximum load on any machine, the load being defined as the sum of the weights of the jobs assigned to the machine. The scheduler also has the option of preempting a job and reassigning it to another machine. Whenever a job is assigned or reassigned to a machine, the on-line algorithm incurs a reassignment cost depending on the job. For arbitrary reassignment costs, we present an on-line algorithm with a competitive ratio of 3.5981 against current load, i.e. the maximum load at any time is less than 3.5981 times the lowest achievable load at that time. Our algorithm also incurs a reassignment cost less than 6.8285 times the cost of assigning all the jobs. This is the first algorithm with a constant bound both on the competitive ratio and on the reassignment factor. For the special cases in which the reassignment costs are either 1 or proportional to the weights, we present several algorithms which improve upon Westbrook's recent 6-competitive algorithm against current load. Our best competitive ratios are 3 + ε and 2 + ε for the unit and proportional cases respectively.

Proceedings of the twenty-fifth annual ACM symposium on Theory of computing - STOC '93, 1993

In this paper we study the problem of on-line allocation of routes to virtual circuits (both point-to-point and multicast) where the goal is to minimize the required bandwidth. We concentrate on the case of permanent virtual circuits (i.e., once a circuit is established, it exists forever), and describe an algorithm that achieves an O(log n) competitive ratio with respect to maximum congestion, where n is the number of nodes in the network. Informally, our results show that instead of knowing all of the future requests, it is su cient to increase the bandwidth of the communication links by an O(log n) factor. We also show that this result is tight, i.e. for any on-line algorithm there exists a scenario in which O(logn) increase in bandwidth is necessary.

Theoretical Computer Science, 1994

The setup for our problem consists of n servers that must complete a set of tasks.

International Journal of Electrical and Computer Engineering (IJECE), 2018

In networks with lot of computation, load balancing gains increasing significance. To offer various resources, services and applications, the ultimate aim is to facilitate the sharing of services and resources on the network over the Internet. A key issue to be focused and addressed in networks with large amount of computation is load balancing. Load is the number of tasks"t" performed by a computation system. The load can be categorized as network load and CPU load. For an efficient load balancing strategy, the process of assigning the load between the nodes should enhance the resource utilization and minimize the computation time. This can be accomplished by a uniform distribution of load of to all the nodes. A Load balancing method should guarantee that, each node in a network performs almost equal amount of work pertinent to their capacity and availability of resources. Relying on task subtraction, this work has presented a pioneering algorithm termed as E-TS (Efficient-Task Subtraction). This algorithm has selected appropriate nodes for each task. The proposed algorithm has improved the utilization of computing resources and has preserved the neutrality in assigning the load to the nodes in the network.

International Journal of Grid and Distributed Computing, 2015

Cloud computing is an internet based technology. This computing paradigm has increased the utility of network where the potentiality of one node can be used by Other node, cloud provides services on demand to distributive resources such as Database, servers, software, infrastructure etc. in pay per use basis, load balancing is One of the unique and important issues for distributing a larger processing lode to smaller processing nodes for increasing total performance of system, in load balancing method the workload not only distribute across multiple computers but also other resources over the network links to gain optimum resource utilization, minimum average response time and avoid overload condition. Different load balancing algorithms have been launched in order to manage the resources of service provider efficiently and effectively. The objective of this paper is to propose efficient scheduling algorithm that can maintain the load balancing and provide improved strategies through efficient job scheduling that would decrease the average response time and increase the availability of more VMs to allocate new jobs from requesting nodes.

Proceedings IEEE International Conference on Cluster Computing CLUSTR-03, 2003

Web application is being challenged to develop methods and techniques for large data processing at optimu m response time. There are technical challenges in dealing with the increasing demand to handle vast traffic on these websites. As number of users" increases, several problems are faced by web servers like bottleneck, delayed response time, load balancing and density of services. The whole traffic cannot reside on a single server and thus there is a fundamental requirement of allocating this huge traffic on mult iple load balanced servers. Distributing requests among servers in the web server clusters is the most important means to address such challenge, especially under intense workloads. In this paper, we propose a new request distribution algorith m for load balancing among web server clusters. The Dynamic Load Balancing among web servers take place based on user"s request and dynamically estimat ing server workload using mult iple parameters like processing and memo ry requirement, expected execution time and various time intervals. Our simulat ion results show that, the proposed method dynamically and efficiently balance the load to scale up the services, calculate average response time, average waiting time and server"s throughput on different web servers. At the end of the paper, we presented an experimentation of running proposed system wh ich proves the proposed algorith m is effic ient in terms of speed of processing, response time, server utilization and cost efficiency.

2007

In this paper, the problem of distributing load of a particular node over m identical nodes of a distributed computing system for minimizing turnaround time is studied first. Then an efficient technique is presented for dynamically scheduling jobs in large-scale, multiuser distributed computing systems that provides a balanced system performance with respect to the scheduling overhead. The nodes are scheduled independently and asynchronously with distinct execution initiation times corresponding to their earliest instant of being overloaded. The technique handles the task of resource management by dividing the nodes of the system into mutually overlapping subsets and thereby a node gets the system state information by querying only a few nodes. The approach is primarily targeted at systems that are composed of general purpose workstation computers having identical processors. Process scheduling decisions are driven by the desire to minimize turnaround time while maintaining fairness among competing applications and minimizing communication overhead. The performance analysis of the technique shows that it significantly reduces the total number of messages required for a node to take scheduling decision.

— Cloud computing represents different ways to design and manage remotely computing devices. Service providers using customized design of cloud computing to public network allows cloud users to establish a relation with cloud computing. There are several heterogeneous nodes in a cloud computing system. Namely, every node has different capability to execute a particular task so only consider the CPU remaining of the node is not enough when a node is chosen to execute a task. Therefore, how to select an efficient node to execute a task is very important in a cloud computing.In this paper, we propose a scheduling algorithm, LBTD, which combines minimum completion time and load balancing strategies. For the case study, LBTD can provide efficient utilization of computing resources and maintain the load balancing in cloud computing environment.

1988

Distributed Computing Systems (DCSs) evolved to provide communication among replicated and physically distributed computers as hardware costs decreased. Interconnecting physically distributed computers allows better communication and improved performance through redistribution (or load balancing) of workload. In this paper, we describe a load balancing strategy for a computer system connected by multiaccess broadcast network. The strategy uses the existing broadcast capability of these networks to implement an efficient search technique for finding stations with the maximum and the minimum workload. The overhead of distributing status information in the proposed strategy is independent of the number of stations. This result is significant because the primary overhead in load balancing lies in the collection of status information. An implementation of the proposed strategy on a network of Sun workstations is presented. It consists of two modules that are executed at all participating computers: the distributed-search module that isolates the maximally and minimally loaded computers, and the job-migration module that places a job based on the load extremes.