Quantitative system performance: computer system analysis using queueing network models
Quantitative system performance: computer system analysis using queueing network models
Adaptive load sharing in homogeneous distributed systems
IEEE Transactions on Software Engineering
A Trace-Driven Simulation Study of Dynamic Load Balancing
IEEE Transactions on Software Engineering
Asymptotic analysis of multiclass closed queueing networks: common bottleneck
Performance Evaluation
HTML: the definitive guide
Client/server programming with Java and CORBA
Client/server programming with Java and CORBA
Asymptotic analysis of multiclass closed queueing networks: multiple bottlenecks
Performance Evaluation
High-performance client/server: a guide to building and managing robust distributed systems
High-performance client/server: a guide to building and managing robust distributed systems
Client/server data access with Java and XML
Client/server data access with Java and XML
Enterprise JavaBeans
Designing Process Replication and Activation: A Quantitative Approach
IEEE Transactions on Software Engineering
Performance Engineering of Software Systems
Performance Engineering of Software Systems
IEEE Transactions on Software Engineering
Designing high-performance software systems: a quantitative approach
Designing high-performance software systems: a quantitative approach
Web-Based System Configuration and Performance Evaluation Using a Knowledge-Based Methodology
EC-WEB '02 Proceedings of the Third International Conference on E-Commerce and Web Technologies
Migrating to Web Services " Latency and Scalability
WSE '02 Proceedings of the Fourth International Workshop on Web Site Evolution (WSE'02)
| Hi-index | 0.00 |
The architects of today's distributed applications have a wide range of Internet technologies, platforms, and design patterns to choose from. In addition to the usual selection criteria of security, portability, maintainability, and cost, performance often determines the selection of one system architecture over another. This paper presents a quantitative technique that can help an architect understand the expected behaviour of an application deployed within a target environment. The technique automatically finds an object allocation that optimizes a performance metric specified by the architect. The technique supports multiple classes of requests and mean response time requirements for multiple workload conditions. Capacity constraints are also considered. These include device utilization limits and the maximum number of customers. A deployed application is described using a Layered Queuing Model (LQM). Non-linear and linear programming techniques are combined with predictive analytic modeling techniques to efficiently compare application configuration alternatives. Both nonasymptotic (no saturated resources) and asymptotic workload conditions are considered.