Convex separable optimization is not much harder than linear optimization
Journal of the ACM (JACM)
Load-balancing heuristics and process behavior
SIGMETRICS '86/PERFORMANCE '86 Proceedings of the 1986 ACM SIGMETRICS joint international conference on Computer performance modelling, measurement and evaluation
Business-oriented resource management policies for e-commerce servers
Performance Evaluation - Special issue on internet performance modelling
Cost-performance optimization of application- and context-aware distributed infrastructures
IEEE Transactions on Systems, Man, and Cybernetics, Part A: Systems and Humans
Optimal service composition via agent-based quality of service
MIWAI'11 Proceedings of the 5th international conference on Multi-Disciplinary Trends in Artificial Intelligence
Capacity planning for vertical search engines: an approach based on coloured petri nets
PETRI NETS'12 Proceedings of the 33rd international conference on Application and Theory of Petri Nets
Towards optimal service composition upon QoS in agent cooperation
International Journal of Computational Science and Engineering
Modelling Search Engines Performance Using Coloured Petri Nets
Fundamenta Informaticae - Application and Theory of Petri Nets and Concurrency, 2012
Hi-index | 0.00 |
Providing quality of service guarantees have become a critical issue during the rapid expansion of the e-Commerce area. We consider the problem of finding the optimal capacity allocation in a clustered Web system environment so as to minimize the cost while providing the end-to-end performance guarantees. In particular, we consider constraints on both the average and the tail distribution of the end-to-end response times. We formulate the problem as a nonlinear program to minimize a convex separable function of the capacity assignment vector. We show that under the mean response time guarantees alone, the solution has a nice geometric interpretation. Various methods to solve the problem are presented in detail. For the problem with tail distribution guarantees, we develop an approximation method to solve the problem. We also derive bounds and show that the solution is asymptotically optimal when the service requirement becomes stringent. Numerical results are presented to further demonstrate the robustness of our solutions under data uncertainty.