Quality of service enforcement for distributed objects

As the number of industrial organisations that adopt the use of shared distributed objects as the vehicle for implementing large software systems increases, new quality issues arise. Four algorithms are presented that are designed to enforce different quality of service criteria for shared distributed objects that have been tested using a previously introduced test generation algorithm. Empirical assessments of these algorithms for three large systems are also presented. These assessments are made in terms of the simulated performance of each system on average loads selected from operational distributions collected during beta release and field use, assuming both Poisson and bursty arrival processes. Heavy loads designed to cause the overall CPU utilisation rates to exceed 90% of capacity were also run.