PERFORMANCE EVALUATION OF TRANSACTION PROCESSING COUPLING ARCHITECTURES FOR HANDLING-SYSTEM DYNAMICS

As the demand for high volume transaction processing grows, coupling m ultiple computing nodes becomes increasingly attractive. This paper pr esents a comparison on the resilience of the performance to system dyn amics of three architectures for transaction processing. In the Shared Nothing (SN) architecture, neither disks nor memory is shared. In the Shared Disk (SD) architecture, all disks are accessible to all nodes while in the Shared Intermediate Memory (SIM) architecture, a shared i ntermediate level of memory is introduced. A transaction processing sy stem needs to be configured with enough capacity to cope with the dyna mic variation of load or with a node failure. Three specific scenarios are considered: 1) a sudden surge in load of one transaction class, 2 ) varying transaction rates for all transaction classes, and 3) failur e of a single processing node. We find that the different architecture s require different amounts of capacity to be reserved to cope with th ese dynamic situations. We further show that the data sharing architec ture, especially in the case with shared intermediate memory, is more resilient to system dynamics and require far less contingency capacity compared to the SN architecture.