No abstract

No abstract

Abstract

ABSTR.ACTThe demand for mobile communications over the past few years has grown at a rapid pace. This demand has led to intensive research and development efforts for complex PCS (personal communication service) networks. Capacity planning and pdrformance modeling is necessary to maintain a high quality of service to the mobile subscriber while minimizing cost to the PCS provider. A question of pragmatic interest concerns the modeling of subscriber or portable movement in a PCS network. Typically, portable movement models are based on a probabilistic distribution function with fixed mean. These models may be an over-simplification of real portable movement patterns. For example, portable movements during rush-hour traffic may be slowed due to traffic jams. However, when the volume of vehicular traffic is less, portables are allowed to move more freely and with greater speed. To capture this type of portable movement behavior, we develop a population dependent PCS model where portable movement is based on the number of portables currently residing in that service area or cell. Because of the large amount of computation required to simulate PCS networks, we implement this model on a distributed Time Warp simulator, which has been shown to reduce the execution time of a single run from 20 hours down to 3.5 hours. Using this simulation model, we study the effect of different call workloads and population dependent portable movement patterns on PCS blocking statistics and present our preliminary results.

In this paper we present the design and implementation of a complex view storage system model that is suitable for execution on a optimistic parallel simulation engine. What is unique over other optimistic systems is that reverse computation as opposed to state-saving is used to support the rollback mechanism. In this model, a hierarchy of view storage servers are connected to an array of client-side local disks. The term view refers to the output or result of a query made on the part of an application that is executing on a client machine. These queries can to be arbitrarily complex and done using SQL. In our performance study of this model, we find that speedups range from 1.5 to over 5 on 4 processors. Super-linear speedups are attributed to a slow memory subsystem and the increased availability of level-1 and level-2 cache when moving to a larger number of processors.