Search

A model is developed for representing computations that arise when deterministic algorithms process workloads whose detailed structure is uncertain. The new computational model, which is based upon the notion of loosely constrained deterministic systems (LCD systems), incorporates a characterization of uncertainty that is expressed entirely in terms of observable properties of the computation itself. This provides a number of advantages over traditional probabilistic models. In addition to providing a more direct link to problems that arise in practical settings, LCD models offer a new perspective on the analysis of risk and support a new method for improving the efficiency of certain computer driven simulations. [ABSTRACT FROM PUBLISHER]

The major resource allocation mechanisms used by the MVS operating system are described, and a queueing network model that represents these mechanisms is presented. The model incorporates a number of refinements that have not been reported previously: the use of shared domains, the explicit representation of separate queues for each domain, and the treatment of domain migration. [ABSTRACT FROM AUTHOR]

Queueing network models have proved to be cost effective tools for analyzing modern computer systems. This tutorial paper presents the basic results using the operational approach, a framework which allows the analyst to test whether each assumption is met in a given system. The early sections describe the nature of queueing network models and their applications for calculating and predicting performance quantities. The basic performance quantities-such as utilizations, mean queue lengths, and mean response times-are defined, and operational relationships among them are derived. Following this, the concept of job flow balance is introduced and used to study asymptotic throughputs and response times. The concepts of state transition balance, one-step behavior, and homogeneity are then used to relate the proportions of time that each system state is occupied to the parameters of job demand and to device characteristics. Efficient methods for computing basic performance quantities are also described. Finally the concept of decomposition is used to simplify analyses by replacing subsystems with equivalent devices. All concepts are illustrated liberally with examples. [ABSTRACT FROM AUTHOR]

The article focuses on writer on computer issues Yonathan Bard's suggestion that distribution-free stochastic results such as Little's law hardly require validation. But distribution-free results express relationships among stochastic parameters and expectations of random variables. Validation involves substituting observed values for these stochastic quantities, or even that they will say obey the same mathematical relationships. Hence, distribution free stochastic results do require validation. In contrast, operational laws such as Little's law express invariant relationships among observable quantities. They require no validation. Second, Bard has noted correctly that the homogeneity assumptions used in certain operational derivation periods. However the same is true of many important models found in science and engineering- e.g. Newton's model of an ideal fluid or Boyle's model of a perfect gas. Such models are useful because their abilities to estimate the behavior of real systems have been well validated.