Exploiting the control power of SDN during the transition from IP to SDN networks

다국어 초록 (Multilingual Abstract)

The emerging software‐defined networking (SDN) paradigm introduces new opportunities to improve network performance due to the flexibility and programmability provided by a logically centralized element named controller. However, a rapid adoption of the full SDN architecture is difficult in the short term due to economic and technical reasons. This paper faces the SDN nodes replacement problem during the transition from traditional IP networks to fully deployed SDN networks. Six different replacement methods are proposed to select the most appropriate set of traditional IP nodes to be upgraded to SDN‐enabled switches at a particular transition stage. To show the effectiveness of the proposed methods, they have been applied on an optimization problem currently studied by the research community: the power consumption problem. An integer linear programming formulation is presented to solve it and a genetic algorithm is evaluated through simulations on realistic network topologies. Results highlight that energy‐efficiency in hybrid IP/SDN networks can be significantly improved by only replacing a reduced number of IP nodes.
This paper faces the SDN nodes replacement problem during the transition from traditional IP networks to fully deployed SDN networks. Six different replacement methods are proposed to select the most appropriate set of traditional IP nodes to be upgraded to SDN‐enabled switches as a function of the optimization problem to be solved. Results of the genetic algorithm proposed highlight that energy efficiency in hybrid IP/SDN networks can be significantly improved by only replacing a reduced number of IP nodes.