Adaptation algorithm for improvement of effectiveness of distributed computer systems

The adaptation of distributed computer systems to changes in their parameters is investigated. The concept of adaptation is based on the correspondence between the control algorithm and the operating system task allocation algorithm on the one hand and the controlled process parameters and those of the distributed computer system hardware and the load on the other. The adaptation to changing parameters and its optimization may be obtained by using the task allocation algorithm. This algorithm, like optimization control programs in multilayer computer control systems, should be activated in discrete moments of time. The problem of trade-off between the cost (overhead) of allocation algorithm computation and the improvement of system effectiveness is discussed. Examples point out the feasibility of this approach.