Panel 2: Policy-enabled Mechanisms for Feature Interactions: Reality, Expectations, Challenges

Thursday, June 12th, 13:15

Abstract. Since the 90's, the problem of feature interactions has been recognized as crucial in designing and implementing telecommunications distributed systems. As distributed systems evolve in terms of features, technology, complexity, and size, the feature interaction problems expand, become distributed and more complex, and require sophisticated solutions. Fundamentally, feature interactions are related to the creation, maintenance, and evolution of new services (telephony, electronic commerce, web services, multimedia, banking, etc.) and to the implementation of these services across distributed, sometimes heterogeneous, platforms. Feature interactions may also refer to software maintainability, software verification and validation, software documentation, and software development process.

While it is advocated that policy-based management systems will become the pillar technology of self-management and autonomic computing systems, problems with the paradigm itself are equally debated in the policy community. For most of existing industrial and academic reported achievements on policy-related aspects, due diligence has to be given before relying on existing techniques, as there is a tendency of oversimplifying the use of the policy paradigm and neglect its own feature interaction problems. Several existing solutions are particularly tailored to a given area (security, intrusion detection, monitoring, etc.), whereas inter-application, inter-network, and inter-feature aspects are more relevant for the use of policies.

The panel will debate emerging technology areas and aspects where there is a great potential for detecting and solving features interactions using the policy paradigm. The guest panelists will bring examples from complementary areas, present their experience on using polices for feature interaction problems, and expose their position on the potential of using the policy-enabled mechanisms for challenging feature interactions problems in telecommunications and software systems. We reckon that the trio of "reality, expectation, and challenges" might express the best we can currently envisage at the convergence of two promising and, at the same time, scrutinized complex paradigms.

The guest panelists might have controversy on their perceived reality, on their short, -mid, and -long term expectation, or might overstate or underestimate the obstacles in combining these two paradigms for achieving practical solutions for feature interaction problems. Hopefully the debate will end up with new avenues for both topics on the academic lane and on the advanced industrial innovation lane, as well.