Modeling can be used to predict the performance of picture archiving and communication system (PACS) configurations under various Ioad conditions at an early design stage. This is important because choices made early in the design of a system can have a significant impact on the performance of the resulting implementation. Because PACS consist of many types of components, it is important to do such evaluations in a modular manner, so that alternative configurations and designs can be easily investigated. Stochastic activity networks (SANs) and reduced base model construction methods can aid in doing this. SANs are a model type particularly suited to the evaluation of systems in which several activities may be in progress concurrently, and each activity may affect the others through the results of its completion. Together with SANs, reduced base model construction methods provide a means to build highly modular models, in which models of particular components can be easily reused. In this article, we investigate the use of SANs and reduced base model construction techniques in evaluating PACS. Construction and solution of the models is done using UltraSAN, a graphic-oriented software tool for model specification, analysis, and simulation. The method is illustrated via the evaluation of a realistically sized PACS for a typical United States hospital of 300 to 400 beds, and the derivation of system response times and component utilizations. developed to offer the advantages of better image accessibility, reduced overhead costs of image handling and improved diagnostic image quality. However, the huge amount of data handling and processing involved in PACS lead to problems related to limited storage and long processing and network delays, which may cause unacceptable response times to image viewing requests. ~ Model-based evaluation can be used to identify such problems and to investigate potential solutions in the early design phase of a project. If successful, such studies can aid in producing better system designs in a shorter period than would be possible if prototyping were the sole evaluation method used. In particular, we can predict the performance of PACS under various conditions by varying different parameter values in the simulation, eg, workloads and hardware specifications. Furthermore, simulation can be used to evaluate the algorithms and software processes that constitute the image processing aspect of PACS, eg, the image compression and image prefetching algorithms. 2 The simulation of these algorithms and processes can then be used asa starting point for their implementation. 1
CopyrightIn this article, we investigate the use of stochastic activity networks (SANs) 3,4 and reduced-base model construction methods 5 in constructing models of PACS. SANs are used to build models of PACS components (ie, modalities, viewing workstations, networks, and database archives), and the reduced base model construction formalista is used to connect these components into different system configurations. This combina...