A New Component Model for Highly Distributed Environements
Résumé
During the last years new distributed platforms have emerged, often qualified as highly distributed environments (HDEs). HDEs still include powerful and robust machines but they are rather composed of resource-constrained and mobile devices such as laptops, tablet computers, personal digital assistants (or PDAs), smart-phones, GPS devices, sensors, etc. Moreover, these devices communicate using variety of dependable and undependable fixed and wireless networks. This fundamental change in the deployment environment was not accompanied by a software model that provides deep understanding and systematic solutions to build compatible software systems. This challenge has been reported as a problem by software developers, software architects, and computer science researchers. This is usually achieved through the utilization of a middleware layer that has for main function (among others) to make remote calls appear as local calls. That is correct for stable distributed systems, however, this same concept, distribution transparency, has been shown to suffer from major shortcomings when applied extensively in HDEs. That leaves us in the following situation: there is excessive and increasing need to build complex mobile and pervasive systems for entertainment and professional uses. And at the same time, the fundamental engineering techniques available are inherited from stable distributed environments, and suffer from several drawbacks and weaknesses when utilized in these new environments. The only available answer currently is applying ad-hoc techniques to overcome these drawbacks and weaknesses. This ongoing Ph.D. work is a direct response to the above mentioned challenge. We are trying to provide a comprehensive answer for the above mentioned challenge. The solution we propose is based on a "paradigm shift" from distribution transparency to localization acknowledgment.