Reverse engineering performance models from trance data

随着教育、卫生、政府和商业组织进入交互计算机网络，分布式计算系统对于信息访问和管理正变得越来越重要。构建这些系统的性能模型已被证明在项目的所有阶段都是有益的。例如，可以使用模型来设置系统组件的性能预算。该模型还可以用来寻找可行的设计和定位潜在的瓶颈。最后，该模型可用于规划已安装系统所需的资源。在所有这些情况下，使用模型可以为对性能敏感的项目节省大量成本。

Distributed computing systems are becoming increasingly important for information access and for administration as organizations in education, health, government, and business move into networks of interacting computers. Constructing performance models of these systems has shown to be beneficial during all stages of a project. For example, a model can be used to set performance budgets for components of a system. The model can also be used to find feasible designs and locate potential bottlenecks. Finally, the model can be used to plan the resources required for installed systems. In all of these case cases, the use of models can lead to substantial cost savings for a performance-sensitive project.