A Statistical Approach to Predicting Workstation Idle Times
نویسنده
چکیده
– Existing networked workstations are becoming increasingly effective as non–dedicated computing clusters. Effective scheduling in such clusters requires knowledge of the availability of the workstations. A successful method for predicting workstation idle periods with statistical distributions is presented. The Weibull probability density function is shown to be useful in modeling workstation availability periods and capable of effective prediction of idle times. Actual workload data is combined with machine logs to compare the Weibull prediction method with an idle–history method, and the results are discussed. Point of Contact: Samuel H. Russ Phone: (601) 325–7775 Fax: (601) 325–7692 Mail: Engineering Research Center P.O.Box 9627 Mississippi State, MS 39762 email: [email protected] Web: http://www.erc.msstate.edu/~russ Key Phrases: Distributed Computing, Network of Workstations Computing, Idle Workstations
منابع مشابه
Optimum Aggregate Inventory for Scheduling Multi-product Single Machine System with Zero Setup Time
In this paper we adopt the common cycle approach to economic lot scheduling problem and minimize the maximum aggregate inventory. We allow the occurrence of the idle times between any two consecutive products and consider limited capital for investment in inventory. We assume the setup times are negligible. To achieve the optimal investment in inventory we first find the idle times which minimi...
متن کاملBudgetary Constraints and Idle Time Allocation in Common-Cycle Production with non-zero Setup Time
Economic lot size scheduling problem (ELSP) for a multi-product single machine system is a classical problem. This paper considers ELSP with budgetary constraint as an important aspect of such systems. In the real world situations the available funds for investment in inventory is limited. By adopting the common cycle time approach to ELSP, we obtain the optimal common cycle which minimizes the...
متن کاملThe Utility of Exploiting Idle Memory for Data-Intensive Computations
In this paper, we examine the utility of exploiting idle memory in workstation pools. We attempt to answer the following questions. First, given a workstation pool, what fraction of the memory can be expected to be idle? This provides an estimate of the opportunity for hosting guest data. Second, what fraction of a individual host' s memory can be expected to be idle? This helps determine the r...
متن کاملAn Idle Compute Cycle Prediction Service for Computational Grids
The utilization of idle compute cycles has been known as most promising and cost-effective way to build a large scale high performance computing system, but not widely used because of the lack of effective idleness prediction techniques. In this paper, we argue PCs at university computer labs have a great potential for the utilization of idle CPU cycles, and propose two techniques for predictin...
متن کاملAn Efficient Algorithm to Allocate Parts to Cells Minimizing Total Tardiness and Idle Times
In the design of a cellular manufacturing system (CMS), one of the important problems is the cell formation in the form of machine grouping and parts family. This paper investigates an allocation of parts to common and specific cells; in such a way that each common cell is able to process all required parts. Further, this paper presents a mathematical programming model comprising such constrain...
متن کامل