Virtual Topology Reconfiguration in Optical Networks by Means of Cognition: Evaluation and Experimental Validation

In optical networking, virtual topologies have been introduced mainly to provide service providers with logical connections equipped with a reserved amount of bandwidth, which can be exploited to interconnect their equipment at the edges of the transport infrastructure. Virtual topologies are thus basically an abstraction of the real substrate, created by means of a process called virtual topology design (VTD). VTD is a complex task, affected by many parameters and constraints, and among them current traffic conditions are very relevant. Indeed, it is possible that after a certain time a virtual topology becomes inappropriate to serve current traffic. In such cases, the virtual topology can be reconfigured by creating new lightpaths or modifying or deleting existing ones, thus possibly creating some service interruptions. In this paper a new virtual topology reconfiguration technique is presented. In this technique, a cognitive entity designs and reconfigures virtual topologies by exploiting traffic forecasting solutions and taking advantage of past history. Moreover, a new transition method is also proposed to reduce the impact of instable routing tables during the reconfiguration process. We demonstrate, by means of simulation, the advantages of the proposed methods, as they reduce both the operational costs and the resources in operation while maintaining low packet loss ratio. Furthermore, we validate the operation of the proposed solutions in an emulated testbed.