Dynamic Load Balancing Based on Applications Global States Monitoring

Abstract

Abstract-The paper presents how to use a special novel distributed program design framework with evolved global control mechanisms to assure processor load balancing during execution of application programs. The new framework supports a programmer with an API and GUI for automated graphical design of program execution control based on global application states monitoring. The framework provides highlevel distributed control primitives at process level and a special control infrastructure for global asynchronous execution control at thread level. Both kinds of control assume observations of current multicore processor performance and communication throughput enabled in the executive distributed system. Methods for designing processor load balancing control based on a system of program and system properties metrics and computational data migration between application executive processes is presented and assessed by experiments with execution of graph representations of distributed programs.

Load Balancing of Java Applications by Forecasting Garbage Collections

“…This prototype was built on top of the Central Directory 5 , which is a lightweight load balancer. This solution was chosen because it is open source and developed in Java, characteristics which facilitated its integration with the MaGC forecast prototype.…”

“…For example, the authors of [11] proposed a technique to estimate the global workload of a load balancer to use this information in the balancing of new workload. Meanwhile, the work on [5] presented a framework for processor load balancing during the execution of application programs. Regarding Java technologies, the authors of [2] enhanced a load balancing algorithm for Java applications by considering the utilisation of the JVM threads, heap and CPU to decide how to distribute the load.…”

PEGASUS DA framework for distributed program execution control based on application global states monitoring

“…A number of papers have analyzed expected advantages of the control design environment based on strongly consistent global states in different applications. This research concerned nature-inspired optimization algorithms including genetic algorithms [6], event-driven parallel simulation [7], and processor load balancing algorithms [8,9]. The interface to design distributed programs is supported by a run-time system for clusters of multicore processors, which provides automated global states construction and respective control design including involved communication.…”