MEMS2014

International Workshop on Modeling and Evaluation of Middleware Systems

Collocated with ACM/IFIP/USENIX International Middleware Conference 2014 December 2014, Bordeaux France

Middleware systems are built using APIs of a middleware platform for integrating systems across networked and heterogeneous computing environments. Existing middleware platforms provide direct runtime support as a container to run distributed software components through a variety of software frameworks and communication protocols. As acceptance of cloud services continues to accelerate, middleware is continuously used to develop solutions for Software-as–a-Service and Platform-as-a-Service. Under such a shift, the requirements on middleware are increasingly focused on Service Level Agreements (SLA), cost, and on-demand resource scaling. As a result, middleware systems need to develop autonomic features such as monitoring the runtime status and being proactive and resilient, which leads to more complexity and uncertainty in the development lifecycle. From software engineering perspectives, modeling techniques and tools help to capture key characteristics of the systems’ behavior and provide early feedback on quality of services. However, existing modeling and evaluation tools are not capable to cover all these aspects. Therefore, new languages, methods and tools are necessary. A combination of universal modeling languages, knowledge management tools such as ontologies and collaborative environments, specialized tools such as agent-based simulators, is necessary. This leads to hybrid models where for instance ontology models of data communications and agent-based behavior models are combined to enhance middleware’s ability of coordinating multiple data sources. It is especially the potential size and complexity (with a variety of features) of middleware systems that constitute a challenge for modeling and evaluation. 

The goal of the workshop is to identify and publish improved methods, tools, techniques, advanced computing capabilities (such as high performance and cloud computing), and standard based open distributed architectures for modeling and evaluating middleware systems.