Chapter 18 PREDICTING GRID RESOURCE PERFORMANCE ONLINE
نویسندگان
چکیده
In this chapter, we describe methods for predicting the performance of Computational Grid resources (machines, networks, storage systems, etc.) using computationally inexpensive statistical techniques. The predictions generated in this manner are intended to support adaptive application scheduling in Grid settings, as well as online fault detection. We describe a mixture-of-experts approach to nonparametric, univariate time-series forecasting, and detail the effectiveness of the approach using example data gathered from “production” (i.e., nonexperimental) Computational Grid installations.
منابع مشابه
Predicting Grid Resource Performance Online
In this paper, we describe methods for predicting the performance of Computational Grid resources (machines, networks, storage systems, etc.) using computationally inexpensive statistical techniques. The predictions generated in this manner are intended to support adaptive application scheduling in Grid settings, and on-line fault detection. We describe a mixture-of-experts approach to non-para...
متن کاملBuilding Online Performance Models of Grid Middleware with Fine-Grained Load-Balancing: A Globus Toolkit Case Study
As Grid computing increasingly enters the commercial domain, performance and Quality of Service (QoS) issues are becoming a major concern. To guarantee that QoS requirements are continuously satisfied, the Grid middleware must be capable of predicting the application performance on the fly when deciding how to distribute the workload among the available resources. One way to achieve this is by ...
متن کاملThe Cost-Based Resource Management in Combination with QoS for Grid Computing
In the grid context, the scheduling can be grouped into two categories: offline scheduling and online scheduling. In the offline scheduling scenario, the sequence of jobs is known in advance, scheduling is based on information about all jobs in the sequence. While, in the online scheduling scenario a job is known only after all predecessors have been scheduled, and a job is scheduled only accor...
متن کاملPerformance Contracts: Predicting and Monitoring Grid Application Behavior
The Internet and the growing availability of inexpensive yet powerful commodity hardware has changed our view of computing. Computational grids have emerged as a new paradigm for high-performance computing. Grids consist of geographically distributed computational resources, such as supercomputers, clusters of workstations, scientific instruments, storage systems, visualization systems and so o...
متن کاملPeer-to-peer protocols and grid services for resource discovery on grids
Resource discovery is a key issue in Grid environments, since applications are usually constructed by composing hardware and software resources that need to be found and selected. Classical approaches to Grid resource discovery, based on centralized or hierarchical approaches, do not guarantee scalability in large-scale, dynamic Grid environments. On the other hand, the Peer-to-Peer (P2P) parad...
متن کامل