MoSiSaS-A Model Driven Methodology For Scientific Software AS A Service

Information technology facilitates the operations of scientific research more and more. However, problems in scientific software development domain also become obstacles and decelerate the progresses of scientific research. These problems are mainly from architecture modeling, mathematic modeling, and execution platform aspects. How to provide the higher quality scientific software with higher productivity and effectiveness becomes an urgent problem.On one hand, Model Driven Architecture is used more and more over the last few years. Central to the MDA philosophy is the concept that an organization should be able to model and build their systems once, and then transition these systems over time as standards and infrastructure technologies evolve and change. On the other hand, Software as a Service (SaaS) provides users on demand technologies and becomes a trend. We believe the advantages of MDA and SaaS can be applicable to the scientific research domains.This research thesis tries to present a methodology called MOSiSaS, which is short for"A Model Driven Methodology for Scientific Software as a Service". We address technical challenges and analyze requirements from specific scientific problem domains, provide guidance in leveraging MDA and SaaS technologies to design and implement scientific infrastructures. MDA is used as a development approach and SaaS is used as a delivery approach in our methodology. We also presents overall architectures and related industry standards of MDA and SaaS, and how they can deliver improvements and project savings in the areas of modeling design, tools support, automated code generation, and software delivery. We use modeling languages like SoaML, SysML and mathematic languages like MATLAB and Modelica with MDA approach in our methodology. An integrated scientific software development tools set is also presented in this thesis. A case study called SiSaS is used here. We hope that our experiences prove to be useful in building similar infrastructures for other scientific applications.