This Industry Viewpoint was contributed by Benny Sands, Director of Business Development - NFV Testing Strategy Lead at QualiTest
We’re off to a great start: both NFV and DevOps are in the title. If you are sensitive to marketing hype, you might be getting a little uncomfortable and we have not even gotten to the end of the first paragraph. I work in a testing company and have little patience for buzzwords or trends, but I keep hearing these abbreviations and feel it’s time to speak up- not with predictions and forecasts but by providing some common sense advice.
For some testing is seen as an afterthought, for others a necessary and others know it simply can’t be overlooked. As a testing company you know where we stand on that. What were are looking at is how testing is going to impact DevOps and NFV both right now and in the future.
A gentle reminder of the promise of DevOps and NFV in case you have missed the hype: DevOps is the new way in which applications are designed, coded, staged and tested. Theoretically, it’s quicker, less buggy, more robust and brings testers, coders, developers and IT resource together so that IT drives the business rather than holds it back.
We have a UK client, giffgaff, embarking on the DevOps process with the aim to provide its customers with features and services quickly and efficiently. giffgaff is a community-based MVNO that cares passionately about its customers and has an almost open-sourced feel of its approach to customer service. The IT resource is working hard to use DevOps so it too can contribute to delivering the sort of customer experience for which it is renowned. It’s DevOps experience is about improving what it can offer to its customers rather than launching new services and the key concern is speed of development and deployement.
Leaving the tech piece for a minute, the key to DevOps implementation is ensuring coders, developers and IT resourcing staff work together to deliver the apps/features that the customer wants. There are clearly plenty of technology challenges but the biggest ones in this case are cultural. Namely, how do we work together effectively? The days of “it’s coded, let’s throw it over the wall for the testers to test” are over. For DevOps to work, everyone in the business must learn to ‘play nicely’, which is a lot easier said than done.
But what about NFV?
So the theory for NFV is that the entire network stack can be virtualised. The benefit to enterprise and operator is that new services can be created and rolled out with a speed that was unimaginable a few years ago. In principle, it means that an operator can monetise its network in ways that its CFO can only dream of; instead of waiting months for a new service, a customer need only wait days. The end game is obvious; operators (and enterprises) spend less on the network but get more from it. It’s a case of wanting your cake, eating it and not putting on weight!
Now for some reality, NFV is very, very complex; there are standards and protocols to adhere to (and agree on), and infrastructure providers and operators that are looking at the opportunity from very different perspectives. Equipment vendors want to sell equipment and operators have a network to maintain and legacy network to look after. A word on legacy- it is more than a word, it’s an investment imperative that keeps any CFO worth his or her title awake at night. The objective is operators have to make the most of what has been spent already before looking at spending more on kit.
But then there is the culture piece to be considered. Historically, network managers monitored what happened and worked with testers to ensure any kit being added to the network performed as ordered. The tests were done in the lab to simulate every scenario so the possibility of a network outage was minimized. Accidents do happen, but proper systems and process reduce the odds. In regards to mindset, the tester and network manage, plan, test, and proceed with caution and all is well on the network and with the end customer. If there are changes to be made, they will happen but it will be done when there is availability.
The trouble is with NFV, DevOps has arrived at the network managers’ door and life has the potential to get very tricky indeed. For DevOps, we think fast, test quick and deploy at speed, which is not what a network manager is used to at all. “Not on my network” might be the mantra to NFV, but if the CFO, CMO, CEO and investors feel that NFV is the key to unlock the network, some serious thinking will need to be done. Gone are the days of a ‘safe’ lab test; now it’s test in on the fly in the cloud. If you are a network manager, I can feel your blood pressure rising at the very thought.
Currently QualiTest is being asked by our operator customers to work out what sort of testing regime is needed for the new services that NFV will be able to offer its customers. We are also working with the likes of Intel and others to test and understand their plans for NFV. But our main effort is to help smooth the path for NFV, working out how to ensure the process works at a functional level and that the technical and business sides of the operator can work together effectively. I am not into predictions or forecasts but I can tell you it’s culture and mindset that holds the key to NFV, as does testing!
If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!Categories: Industry Viewpoint · NFV · Software