METHOD FOR ESTIMATING THE PERFORMANCE OF LARGE INFORMATION SYSTEMS WITH TRANSACTION CLUSTERING
Annotation
A method for estimating the upper and lower bounds of large information system performance with the account for the system data integrity requirement, is proposed. The integrity is ensured by locking the computing resources necessary for the transaction execution and releasing them when the transaction is completed or cancelled. Clustering of transactions allows for parallel processing of different user requests belonging to different clusters. The features of query processing give no way of analytical assessment of large information systems performance, and a full-scale or simulation experiment is time consuming. The model of a large information system is formalized in the form of a mass service network. The complete set of routes in the queuing network is given by the number of clusters of similar transactions. The performance is estimated by the system response time.
Keywords
Постоянный URL
Articles in current issue
- APPLICATION OF METHODS OF FRACTAL IMAGES ANALYSIS IN SOLVING ENVIRONMENTAL ASSESSMENT AND OBJECT RECOGNITION PROBLEMS
- MODEL FOR ASSESSING THE READINESS OF COMPLEX TECHNICAL SYSTEMS TAKING INTO ACCOUNT THE METROLOGICAL CHARACTERISTICS OF CONTROL TOOLS
- ALGORITHM FOR PREDICTING THE STATE OF EQUIPMENT BASED ON MACHINE LEARNING
- COORDINATED WORK OF TWO MOBILE ROBOTS WHEN PERFORMING OPERATIONS AUTOMATICALLY IN A COMPLEX ENVIRONMENT
- PROMISING METHODS FOR MEASURING THE THERMAL CONDUCTIVITY OF SOLIDS
- MODEL OF SPATIAL-TEMPORAL SYNTHESIS OF THE COMBINED APERTURE FUNCTION OF SPACED ANTENNA ARRAYS
- CALCULATION OF OBSERVATION PERSISTENCE DURING RADAR SURVEY