Model-driven engineering for distributed, reliable, adaptive, web-based software

点击翻译按钮获取中文摘要

Expected future economic, industrial and societal innovations place significant challenges on software and its underlying development processes. For instance, the software required for many of the envisioned Internet of Things or smart infrastructure applications will have to combine several characteristics to an unprecedented degree, including scale, distribution, heterogeneity, flexibility, evolvability, reliability, safety, and security. According to experts, current software engineering techniques and tools are insufficient to build software of this level of complexity effectively and efficiently. The proposed research will address this problem with a particular focus on distributed software with high robustness, reliability, and adaptation requirements. Robustness and reliability are basic quality attributes required by most software. Adaptation extends previously articulated visions of autonomic computing and facilitates, e.g., the use of software under changing or unknown operating conditions or usage patterns. The work will combine the state-of-the-art in distributed systems with concepts, notations, techniques and tools from fields that promote abstraction and automation including model-driven engineering (MDE), software architecture, and domain-specific languages, all of which have the proven potential to help control complexity and allow for the efficient, effective development of complex, high-quality software. The work will also leverage recent applications and extensions of the actor model. Despite its simplicity, this architectural style of organizing concurrent and distributed systems has recently seen very successful use for a range of large-scale, industrial distributed applications, including high-volume telecom systems and online games such as Halo on Microsoft's Xbox. Actors thus appear to be an ideal candidate to provide a suitable foundation to our work. Since many future distributed systems will be web-based, the research will also build on the state-of-the-art in web technology and extend it with better support for the design of large-scale, robust, and reliable distributed software. In sum, the proposed research addresses an important, timely problem in a novel way. Drawing on results from a range of fields in computer science, it will advance the state-of-the-art in the development of software that is distributed and adaptive, yet also robust and reliable. It will help Canadian industry realize the innovation potential of the Internet of Things (IoT) by helping them build the software behind many envisioned IoT applications. It will provide much needed support for the use of current web technology and the JavaScript ecosystem to design and build reliable large-scale systems. Moreover, it will provide training for a broad range of students (3 PhD, 8 MSc, and 5 BSc) and give them the knowledge and skills they need to become sought-after graduates and make valuable contributions to Canadian industry and academia.