Cognitive Decision Entities for Cooperative Management in Networks of the Future (CoDECoM) is a collaboration project between DAI-Labor and Huawei, funded by Huawei’s Innovation Research Program (HIRP). CoDECoM targets an efficient implementation of a knowledge plane based on distributed cognitive nodes and a distributed decision making solution, to address the problem of lacking efficient methods to re-configure the network for a highly dynamic environment. The project intends to design cognitive agent and converged control/management agents and to evaluate cognitive functions in the network nodes that collaborate to provide dynamic management (operation) and resource optimizations based on operator policies.
Within this research direction, CoDECoM proposes:
- A new cognitive node architecture consisting of two sub-components – The cognitive agent takes part in a distributed control process with all other agents on the end-to-end service path. It incorporates the network operator policies in making optimal decisions. The converged control/management agent implements reconfigurable routing based on the decisions of the cognitive agent.
- An efficient implementation of the knowledge plane based on a distributed peer-to-peer storage on individual network entities – Access to the knowledge plane will be through an efficient implementation of publish-subscribe paradigm. The organization of knowledge, as well as the subscription to contents will be through low-weight ontologies, which allow efficient establishment of meaning across different domains.
- A network-application interface gateway – An important part of the knowledge plane information comes from the applications. Furthermore it is important for the applications to describe their requirements to the network entities. CoDECoM’s simple network API will serve for this purpose.
- A testbed on which diverse application scenarios can be tested – The node architecture and integration of CoDECoM components will be tested on a proof-of-concept demonstrator.