This report investigates the critical importance of automation to the success of virtualized telecom networks and operations. In earlier research, ARG has documented that virtualization changes the fundamental economics of networks and their operation. We believe that unless operational assumptions change and automation/autonomic methods are widely adopted, the multi-billion dollar investment in NFV, SDN and Cloud will be severely compromised.
Until recently, automation within the network has been largely impractical, limited by practical considerations. However, as networks evolve from physical instances to virtual instances, automation is becoming not only feasible – but as such, necessary in order to be competitive in terms of agility and cost.
In this report we cover what it takes to implement successful, efficient, lasting automation. We begin with “first principles” – looking at Control Theory and learning’s from other industries. We apply that learning to emerging virtualized networks, hybrid networks (reality) and the management stack as it must evolve. We conclude with best practices.
This report includes the roles of analytics, assurance, orchestration, policy and of new optimization algorithms. It concludes with a “reference architecture” for closed loop automation, and with a set of business ad technical objectives that can be used to evaluate the industry’s progress