Provisioning Legacy Simulation Applications in Product Lifecycle Management via a Cloud Platform

Provisioning legacy simulation applications for Product Lifecycle Management (PLM) is very challenging. There are two key issues: (i) Hosting applications for simulation job executions. These legacy applications are developed in various platforms, usually monolithic, and may heavily consume system resources in the runtime. (ii) Input/output data storage. During the lifecycle of a project, huge amount of design data is generated by various design tools. When a design is completed, it needs to be validated by numerous simulation applications, which consequently create substantial amount of result data. Therefore, among others, the scalability in both system runtime resources and data storage is one of the biggest challenges. A generic cloud solution, which promises elastic scalability in both computation resources and data storage but lack of coordination between runtime resource allocation and data placement, is insufficient for supporting these legacy applications efficiently. In this paper, we share our experience on implementing a cloud solution that incorporates both scalable application hosting and usage-aware data placement into an integrated mechanism. With such a mechanism, our system can not only scale out according to workload, but also improve performance by increasing the possibility of local data access.