Understanding Scalability Issues for a Distributed Simulation Environment Using Intelligent Coordinated Entities.
April Renee Crockett, Rodger Maarfi, Srini Ramaswamy, Eric L. Brown, Mike Rogers · 2005
Distributed applications are no longer an exception to software architecture design. Scalability is a strong motivation of many organizations to implement a distributed application. This paper presents a scalability case study using a distributed highway simulation environment. We report on improving an application’s scalability using different measures related to the simulation architecture. Specifically, we study thread reorganization and network communication reorganization issues and their effect on scalability. We use performance and efficiency as the main tests for application scalability considerations. We use CPU overheads (measured in terms of CPU utilization), collision detection, thread sleep cycle, and network efficiency as the measures for testing performance and efficiency. We adopt a scheme with successive modifications, so that we can learn from each individual effort to improve scalability and judge collectively improvements. 1.