Seamless Connectivity and Middleware
The vision driving ARTEMIS is a major evolution of our society in which most systems, machines, and even ordinary objects will be transformed into digital, communicating, information processing, self managed resources. These transformations will be possible through advances in embedded systems technology and its largescale deployment, not only in industries and services, but all areas of human activity. “Seamless Connectivity & Middleware” along with “Reference Designs and Architectures” and “Design Methods & Tools” been identified as the three main technological chapters that should be developed by ARTEMIS to support the development of high value-added embedded systems. Reference Designs and Architecture describes common requirements and constraints that should or could be taken into account for future embedded systems, and prepares the explicit issuing of designs and architectures application domains. Design Methods and Tools asserts how these systems are going to be designed in the future, in order to achieve a number of conflicting goal: system adequacy to requirements, system optimization, design productivity, and eventually time to market.
The aim of Seamless Connectivity and Middleware
However, embedded systems will not exist if they are not embedded into products.This means in the future being part of wider distributed systems, or being themselves distributed systems. In a more or less dynamic way, systems will be able to join an already existing population of communicating systems, and to become consumers, providers, or part of services to users. So, markets will not open up if basic technology implementing common standards are not available for making seamless connectivity possible. Seamless connectivity means the possibility to exchange at the physical level - networks - at the logical level - data - and at the semantic level - information, and knowledge. Moreover, embedded systems must be organized or must organize themselves to achieve generic functions, and be able to do so under constraints and various events, e.g. they must be safe, reliable, must make diagnosis and maintenance easy. Once networking technology is available - and wireless networking will play a major role in the future - all this is ensured by adequate middleware. So, Seamless Connectivity and Middleware is at the heart of the future embedded system markets, and for products including embedded systems. In fact, if ARTEMIS can provide Europe with adequate technology and standards in this area, new yet unforeseen possibilities of application will be made possible, and this will create a rich environment for creativity, innovation, and eventually growth, and high skilled jobs. It is not exaggerated to say that the impact will be even stronger than the Internet almost 15 years ago.
Domain clustering
However, if we look at the versatility and the complexity of application domains, which are grouped within ARTEMIS in the four application domains: Industrial Systems, Nomadic Environments, Private Spaces, Public Infrastructure, it is clear that this is not a simple, straightforward task. Indeed, it seems that there is not a single middleware design that can fit requirements from all those domains. Moreover, in the current situation, specific, partial technology exists, and is more or less deployed in application domains, and partially standardized in standards specific to application domains. However, one feels that technology and standard development effort could be shared among domains; more importantly, technology should sometimes be shared, because it corresponds to products that will connect; there even exist cases where products will be developed only if some common infrastructure is shared among markets that are separate today. In this case, common technology and standards is absolutely mandatory. This leads to domain clustering. Domain clustering aims at exhibiting a limited number of “abstract domains”, that share common characteristics of (part of) the application domains. Some clusters correspond to “could share” characteristics, others to “should share”. Domain intersections are also stressed.
Method of work
This is what the working group preparing this Strategic Research Agenda synthesized.Application domains were examined separately, and experts from the corresponding industry tried to express the kind of system that they expect to become products in the future.This analysis is only partial and incomplete, because the future is obviously blurred, and because it is not possible to be too precise in future product development. Industrial experts were asked to express when possible future systems in a simple, common “qualitative” vocabulary, that we called a multidomain abstraction.This aims at making domain clustering easier.
If the analysis is correct, then domain clusters show important characteristics that:
- call for developing a common middleware within ARTEMIS for this cluster
- from R&D in attached Research Priorities
- towards technology standardization.
The last step consists of defining those research priorities. Research Priorities are presented, and were also discussed, according to their scientific and technological intrinsic coherence.They are put in coherence with the former analysis, by showing how they contribute, or do not contribute to clusters. Indeed, the process driving to this document was not that straightforward.We started from several points at the same time, and through a series of interactions between topics, we converged on the current version.
How this SRA should be used
It is clear that the analysis that led to the present document was necessarily limited in time and resources, and that future work will be required to confirm or refute the conclusions, to check and refine them. However, these conclusions will not become reality unless industry from various sectors agrees to cooperate on some cluster development goals. It seems essential for ARTEMIS to strongly support discussion between industry from various domains, together with researchers, to develop in a cumulative and efficient way common middleware technology and standards.This is the very essence of ARTEMIS. Once consensus on domain clustering is reached, the final issue concerns the strategy for ensuring success of technology on markets.Various strategies can be pursued, from open source development to selling proprietary middleware implementation.This discussion could not be undertaken by the group within its limited lifetime. It remains to be conducted in a consolidation process with the other SRA groups, and innovation group, and even probably beyond.
Working Groups have indentified research priorities for this domain, see the documents:
Documents
