Holistic integrated decisions trade-off baseline for an optimal virtual IP multimedia subsystem (IMS) architecture

Abstract

Network Functions Virtualization (NFV) reveals the challenge of transformation from traditional monolithic to optimal virtual architectures, while meeting the standards-driven and diverse stakeholders functional, performance constraints and conflicting objectives: i.e. maximizing flexibility (scaling, growth), minimizing expense in terms of financial and human effort and intervention (CapEx, OpEx), etc. We case study the IP Multimedia Subsystem, as being the most complex and important NFV instance. Our approach, specifying a customizable baseline to arbitrate and to subsequently optimize between the NFV strategic objectives and enablers, defines an optimal functional architecture adapted to the context at any level and scope: from a specific virtual network function resources (virtual machines) till infrastructural (multiple virtual IMS systems) requirements. To trade-off for the optimal functional options, we explore the multi-objective optimization method, incorporated with the data gathered from an adapted Model-Based Architectural Framework, PESTEL and FURPSE analyses. We further discuss our vision: self-regulating system and the related constraints.