TIAS 2013 - 1st International Workshop on Trust in Autonomic Systems (TIAS)

Autonomic systems are envisaged to enhance future networks by moving from manual management to dynamic and autonomic control and management. In particular, autonomic systems determine their behaviour, in a reactive or proactive manner, based on the external stimuli (environment aspects), as well as their goals, principles and capabilities while cognitive systems also introduce the sense of experience and knowledge. In the case of cognitive networks, this definition can be translated as the ability to dynamically select the network’s configuration, through self-management functionality that reaches optimal decisions, taking into account the context of operation (environment requirements and characteristics), goals and policies (corresponding to principles), profiles (corresponding to capabilities), and machine learning (for managing and exploiting knowledge and experience). However, autonomicity and thus lack of human intervention in network management raises concerns on how trustworthy such networks can be and how to build the road to their wide adoption and deployment. Those concerns come from all ITC actors while manufactures and academics are trying to address such issues so as to move forward with autonomic solutions.
From the operators’ point of view stability issues, reliability of the mechanism, performance, alternatives if the system fails and of course economical aspects in terms of how OPEX and/or CAPEX will be affected by moving to autonomic solutions consist the main components of trust. Even if manufacturers move forward, if operators do not trust autonomic solutions they will not exploit them. For example, multiples SONs are documented in STD and available, but their deployment could be slowed, or even blocked by the coordination issues which need to be dealt.