A generic cross-layer architecture for autonomic network management with network wide knowledge

Technology advances and services of modern networks have noticeably complicated their management such that in the near future it will be almost unfeasible for human administrators to monitor them. To manage this complexity, IBM has proposed the autonomic paradigm with its four self-management properties. Autonomic computing/networking intends to design computing architecture capable to self-adapt its configuration and to self-optimize its performance depending on its state in order to fulfil high-level goals predefined by the human operator. In this research work, we introduce AutoXL, a cross-layer autonomic architecture. AutoXL adopts a Knowledge plane and a Policy plane. The Knowledge plane is responsible of the management of knowledge (local and network-wide) which enables an autonomic element to be self-aware of its environment. The Policy plane is responsible of managing policies guiding the autonomic element to achieve the predefined objectives. Furthermore, we propose guidelines for autonomic architectures evaluation.