Telcos are facing an unprecedented amount of change in their business environment, in customer expectations and as a result of introducing virtualization to their networks. They want to be able to adapt in a lean and agile manner, understanding what needs to change in their organization optimally to meet those goals and then driving those changes in a disciplined and automated way. Business-driven optimization is the adaptation mechanism that will help operators survive an era of rapid and unpredictable change. It is the practice of determining the business goals behind the change, finding the best solution to meet them and ensuring that the objectives are achieved. In a telecom context, a business goal might be the most profitable way to deliver a service to a particular customer or the most profitable way of deploying a network.
Business-driven optimization is not a one-time activity, however. Due to the rapid pace of environmental change, initial solutions can suffer from entropy, a challenge that is compounded by the dynamic nature of a virtualized network. So optimization must be carried out continually, adjusting for and adapting to any changes that come along. Nor is it sufficient for operators to optimize brilliantly in silos – for a single layer of the network or one specific type of service, for example. In a complex world with competing pressures and priorities, business-driven optimization must be able to take into account multiple businesses, technical and regulatory variables across the entire network and service landscape. This is a massively demanding task but one that reflects the reality of operating a sophisticated business today.
Given the pace and scale of the optimizations that operators will need to apply, business-driven optimization will need to be highly automated. Operators are already exploring the automation needed if the virtualized network is to deliver the agility and cost benefits promised for it. In the automation scenario for the virtualized network, business-driven optimization is the “brain,” analyzing input from the virtualized environment, modelling the input against business objectives, generating a new optimization solution and triggering other automation tools to make the required changes.
The fact that the optimization process needs to be fast, holistic and continuous not only implies automation but also that it should be driven by artificial intelligence and machine-learning techniques. Business-driven optimization powered by artificial intelligence (AI) can cope with multi-variable complexity at speeds impossible to achieve with conventional optimization tools. Machine-learning techniques enable business-driven optimization to acquire and process knowledge without the need for coding, making it highly agile and responsive to change.
This paper describes a business-driven optimization approach and why it’s different from the optimizations that operators carry out today. It assesses the features and benefits of this approach and explains the key requirements for a centralized optimization platform that supports it.
