Semi-distributed load balancing for massively parallel multicomputer systems

Abstract

This paper presents a semi-distributed approach, for load balancing in large parallel and distributed systems, which is different from the conventional centralized and fully distributed approaches. The proposed strategy uses a tw<rlevel hierarchical control by partitioning the interconnection structure of a distributed or multiprocessor system into independent symmetric regions (spheres) centered at some control points. The central points, called schedulers, optimally schedule tasks within their spheres and maintain state information with low overhead. We consider interconnection structures belonging to a number of families of distance transitive graphs for evaluation, and using their algebraic characteristics, show that identification of spheres and their scheduling points is, in general, an NPcomplete problem. An efficient solution for this problem is presented by making an exclusive use of a combinatorial structure known as the Hadamard Matrix. Pedormance of the proposed strategy has been evaluated and compared with an efficient fully distributed strategy, through an extensive simulation study. In addition to yielding high pedormance in terms of response time and better resource utilization, the proposed strategy incurs less overhead in terms of control messages. It is also shown to be less sensitive to the communication delay of the underlying network.

Load Balancing in Cloud Computing: A State of the Art Survey

“…In addition, in this approach all nodes can communicate with each other for achieving a global goal in the system which is called cooperative or every node can work independently for just achieving a local goal that is noncooperative form. But, in a non-distributed scheme [40][41][42][43], the responsibility of balancing the system workload would not be performed by all system nodes. In centralized approach in non-distributed scheme; a single node only can execute the load balancing mechanism among all nodes.…”

Hash-based proximity clustering for efficient load balancing in heterogeneous DHT networks

“…In contrast to the "virtual server" approach, the item moving scheme keeps P2P network scalability and efficiency. Ahmad and Ghafoor [2] proposed a semi-distributed approach for load balancing in large parallel and distributed systems. This method uses a two-level hierarchical control by partitioning the interconnection structure of a distributed or multiprocessor system into independent symmetric spheres centered at schedulers.…”

An Approach on Semi-Distributed Load Balancing Algorithm for Cloud Computing System

“…Node continuously interacts with each other which generate more message than non-distributed in the dynamic load balancing algorithms in distributed nature [4]. As the message is transmitted among the node for the interchange of system update it may lead to cause to system stress and may lead to affect system performance [11].…”

“…Central node is responsible to execute the load balancing algorithm [4]. In semi-distributed form a cluster is formed by a group of the nodes of the system, load balancing is happening in each cluster is of centralized form.…”

“…Among the node in the cluster a central node initializes a to take care load balancing within that cluster. Semi distributed take more message as compared with centralized dynamic load balancing to reach a decision however centralized dynamic load balancing is more suited for network with small size which is not good for large network because load balancing is controlled by a central node if anything happen to the central node than whole network may fail or system may cause delay in processing the task [4] [7] [8].…”