“Automation” is rising back up the agenda for service providers.

In part, this is due to continued squeeze on costs, lack of enough specialist resources, and so on.

However it also reflects a dawning realisation about virtualization in general, and NFV in particular: automation is the whole point.

According to conversations with HeavyReading, over half of telcos now regard “a high degree of automation” as integral to delivering on the business case for NFV.

But in a virtual world, automation ain’t quite what it used to be. HeavyReading’s Head of Cloud and NFV illustrated the point neatly on a recent webinar.

Automating the instantiation of VNFs is a pretty obvious thing to want to do. But the range of different factors involved demonstrates the difficulty:

LightReading VNF Placement
Courtesy of LightReading

Understandably, service experience is the most critical factor to get right.

However, cost is high in the mix, and regulatory constraints were cited by 45% to be a Critical or Important factor. So when we talk about “automation”, we’re now talking about a much wider mix of business and technical decision-making. Spinning up a virtual network function is, frankly, the easy part.  The hard part is figuring out whether you even should.

The more enlightened telcos are recognizing that truly radical change is the target to aim for in today’s market. Rather than looking just for gradual, incremental improvements and a few % off here and there, service providers must now set their sights on radically lower costs, and higher profitability.

Automation is certainly part of the answer – but in an interpretation that business, technical and regulatory factors must all be combined within one single operation.

You can listen to more of the commentary from with HeavyReading’s Principal Analyst for Cloud and NFV here.