Automation was a key point of discussion at Layer 123 event in The Hague last week.
In some ways it reminds me of other terms the industry has used that means different things to different people. Remember “customer experience management” and “analytics”? – What are we talking about specifically and where do we accrue the business benefits? One way to bring clarity to the topic of automation is to look at a framework for how it might work.
Closed-loop control systems have been applied to many industries. A closed loop consists of a controller and sensor. In essence, we measure inputs (network events) which acts as a sensor and compare these events to the desired state (service). Automation can’t exist without understanding and applying actions. The understanding is another area that the industry community is seeking to improve and advance.
This (generating true understanding) is where Appledore Research Group expect to see value creation. We have the compute power now to process large streams of data. The innovation will come from robust machine learning models. This is the ability to understand the state of the service and “tune” the feedback loop via the controller (orchestrator) in fast cycle times where human intervention either falls short or is unable to scale to the requirements of robust cloud services.
Currently, we are looking for the evidence where CSPs and suppliers are moving forward in a step function. I think it will materialize in nested loops at the domain level and for specific services. We already have evidence where SON and SDN networks have led the way. Where might we expect to see it in the NFV domain? More importantly how will the industry apply it across technology domains and geographical boundaries?
Patrick Kelly, Founder and Principle Analyst
our published Framework on automation and control theory is here:
and our “sister” publication on how Service Assurance fits in is here: