Internet based monitoring and control of embedded systems (EES.5413)

The proposed project will advance the usage of the Internet and of Internet technology for monitoring, management and control of embedded systems in a variety of application domains (represented by Philips for the telephony and medical domain, TNO Industrie and several SME companies in the medical, communication and machine control domain) [SME = Small and Medium-sized Enterprises]. We will develop innovative software architecture concepts to resolve a number of issues that presently constitute major obstacles for taking full advantage of the Internet and of Internet technology, viz. security requirements, conceptual integrity requirements, real-time constraints, and resource constraints. The industry partners will provide the requirements with respect to the particular application domains. The university partner uses this input to work on the necessary theory and methods and the monitoring and control channel.

We will develop an experimental software architecture, including several versions of a distributed interaction object which will be called a monitoring and control channel and which is expected to full the above mentioned requirements. We will develop the necessary theory and a prototype implementation.

The monitoring and control channel is a middle-ware solution which will be built on top of existing protocols. The design of the monitoring and control channel will be a combination of several successful concepts from recent distributed systems research, viz. the concept of interaction object, the concept of soft state, and data-push architectures. The monitoring and control channel will hide most of the complexities of the Internet protocols and hide the security issues from the user and from the application programs. This will reduce the complexity of the application programs. We will apply advanced formal methods to the specification and design of the monitoring and control channel.

Demonstrators will be built for one or more non-trivial application domains by the joint eorts of the university and the industrial partners (two initial demonstrators have been built already by EESI and experience with traditional architectures, one based on HTTP/CGI and another on UDP/Java, has been gathered). The industrial partners will validate the concepts in several application domains. Moreover the project will offer a framework for the exchange of ideas, needs, experiments and experiences amongst the partners.

It is a misunderstanding that the research questions of web-control will be solved in say half a year from now, just because working demonstrators exist. Problems such as e.g. security are not solved at all (as an example, think of software download in relation to the problem of aggressive viruses) and moreover, the costs of terminals and the \\problem of the last meter" need to be taken seriously, which implies that really new architectures are needed (except perhaps for a restricted class of professional systems which can afford to put a complete workstation inside the embedded system).

For results, see http://www.win.tue.nl/~johanl/projects/EES5413/

Five companies are involved in this project.