Algorithm Design for Computer System Design

For a long time the design of data processing systems has been mainly based on experience and practical considerations more than on formal quantitative approaches. This was true both in the design of computer architecture and in the design of software systems such as operating systems and database systems. Two relevant exceptions were the early studies in switching theory, concerned with such problems as minimization and reliability, on the one hand, and on the other hand, the general mathematical approach to computer system modeling and performance evaluation. More recently, the evaluation of distributed computing related to technological advances in microelectronics, has increased the need for quantitative studies for optimizing the design of computer systems.
As the complexity of computer systems grows, the need for formalization and theoretical analysis is becoming more and more important. The development of formal semantics has provided tools for dealing with correctness and other desirable properties of distributed computing, the development of formal models in different areas (such as distributed system layout, data base design, computer network topology, scheduling and routing) has provided tools for dealing with efficiency and performance optimization; advances in theory of algorithms design and technological increases in computing power have led to the feasibility of the exact or well approximated solution of large scale optimization problems; finally advances in the theory of computing and analysis of algorithms and data structures have led to a new approach to the design of algorithms for tbe efficient solution of bard problems related to distributed processing systems. Such problems as optimal memory management, optimal design of computer networks and multiprocessor systems, optimal layout of VLSI systems, efficient exploitation of parallel computing systems, optimal management of database schemes, concurrency control, bave been thoroughly investigated recently with relevant practical results.