Impact of Fluctuating Goals on Adaptability of Evolvable VNF Placement Method

Abstract

Software Defined Network (SDN) and Network Function Virtualization (NFV) are effective techniques to deal with dynamically changing network environments. Furthermore, the combination of SDN and NFV permits telecommunication service providers to offer sequences of virtualized network functions to their users through Service Function Chaining (SFC). In the context of SFC, the Virtual Network Function (VNF) placement problem, i.e., determining where the virtual functions should be located in the network, needs to be solved dynamically whenever new function chains are requested. In our previous work, we proposed an evolutionary method for dynamic VNF placement problems named Evolvable VNF Placement (EvoVNFP). This current paper aims at evaluating EvoVNFP in greater detail to clarify the influence of the parameter settings on the performance of EvoVNFP. Results from computer simulations show that appropriate settings of sub-goal period lengths and number of mutations help improve the adaptability and convergence speed of EvoVNFP.