Quantitative system performance: computer system analysis using queueing network models
Quantitative system performance: computer system analysis using queueing network models
A Note on the Computational Cost of the Linearizer Algorithm for Queueing Networks
IEEE Transactions on Computers
Object-oriented development: the fusion method
Object-oriented development: the fusion method
OSF DCE: guide to developing distributed applications
OSF DCE: guide to developing distributed applications
IEEE Transactions on Computers
COBRA fundamentals and programming
COBRA fundamentals and programming
Linearizer: a heuristic algorithm for queueing network models of computing systems
Communications of the ACM
Presenting Java
Performance Engineering of Software Systems
Performance Engineering of Software Systems
HTML 3.2 and CGI Unleashed
Performance Evaluation of Client-Server Systems
IEEE Transactions on Parallel and Distributed Systems
IEEE Transactions on Software Engineering
Automatic Generation of a Software Performance Model Using an Object-Oriented Prototype
MASCOTS '95 Proceedings of the 3rd International Workshop on Modeling, Analysis, and Simulation of Computer and Telecommunication Systems
Hi-index | 0.00 |
Recent advances in distributed object and Internet technologies have made it attractive for organizations to distribute application functions. Typical projects include: the re-hosting of legacy applications that move application functionality to or from mainframe/server environments, the creation of new target independent interfaces for legacy systems, and the development of new applications altogether. Design concerns for such systems include security, reliability, and performance. The performance of these systems often defy intuition and must be taken into account during their design. In this paper we present a performance engineering tool for developing predictive models for such systems. The tool automates model construction by developing the structure of the model and measuring parameters that are difficult to estimate or capture manually. Designers can then focus on the performance impact of system configuration alternatives. We show how these results have been integrated into a prototype of IBM's Distributed Application Development Toolkit (DADT). A case study is presented that considers the hosting of sample application across three architectural models: Client/Server using DCE, Web Server/Server using HTML/HTTP/CGI, and a JAVA/CORBA-ORB/ Server model.