|
Abstract : |
Load balancing is often used to ensure that nodes in a distributed systems are equally loaded. In this paper, we show thatfor real-time systems, load balancing is not desirable. In particular, we propose a new load-pro ling strategy that allows the nodes of a distributed system to be unequally loaded. Using load pro ling, the system attempts to distribute the load amongst its nodes so as to maximize the chances of nding a node that would satisfy the computational needs of incoming real-time tasks. To that end, we describe and evaluate a distributed load-pro ling protocol for dynamically scheduling time-constrained tasks in a loosely-coupled distributed environment. When a task is submitted to a node, the scheduling software tries to schedule the task locally so as to meet its deadline. If that is not feasible, it tries to locate another node where this could be done with a high probability of success, while attempting to maintain an overall load pro le for the system. Nodes in the system inform each other about their state using a combination of multicasting and gossiping. The performance of the proposed protocol is evaluated via simulation, and is contrasted to other dynamic scheduling protocols for real-time distributed systems. Based on our ndings, we argue that keeping a diverse availability pro le and using passive bidding (through gossiping) are both advantageous to distributed scheduling for real-time systems., |
