Helping operators to guide virtualisation into a reality

Published on:

TM Forum has released a report highlighting the challenges facing operators as they attempt to bring virtualisation further into an operational environment. GTB speaks to Barry Graham, the industry organisation’s director of product management

What have been some of the initial lessons from NFV deployments?

One of the most interesting examples I heard was where an operator had looked at addressing the organisational aspects of NFV with an ‘island’ deployment. As well as implementing a solution, they also created a complete and standalone cross-functional team to manage it. So, in addition to getting valuable experience out of the technology, they also got valuable experience surrounding new organisational aspects, new ways of working, and new processes that they can now leverage as they transform the rest of the organisation. 

What are the benefits and limitations of experimenting with NFV in an ‘island’ deployment? 

The benefits are clear: operators are able to get experience and value from NFV at much lower risk. However, the business benefits that can be derived from these islands tend to be limited as well. Typically, the island is only able to replicate an existing service. We know from our survey that business flexibility and agility are the things that are really driving NFV, and we know from several TM Forum Catalyst projects that automated dynamic behaviour really is possible and can deliver those benefits, but it typically needs a good tie-in to the rest of the support systems to produce value.

What are the next steps for carriers to bring NFV into an operational environment?

We can see from our survey that adoption really is picking up now. From our conversations with members, we understand that the deployments today are either being rolled out on a very limited scale, or are situations where a particular function or element is virtualised, but is only being used as if it were a physical one. An example of this is vCPE deployments. 

Typically, where operators have done this, they are providing the same functionality that would have run on the CPE, but hosted in the cloud. We think this will continue for some time as the amount of virtual infrastructure is slowly growing. 

Then, operators need to start actually operating it as virtual infrastructure and start to exploit its full power. Our vCPE Catalyst actually followed this same path – they proved core functionality in the first phase, and then started looking at how they could use the power of virtualisation to enhance the customer experience. 

Can you elaborate further on some of the key strategies highlighted in the report? 

Just after our report was published, we held one of our digital leadership summits focused on NFV, where we had representatives from about 20 service providers discussing some of these key points. What was clear was that the organisational and cultural aspects really need to be addressed. Getting the right ownership and team leadership is key to ensuring that the operationalisation of NFV is not seen as a networks team activity, but that the impact across the whole business is addressed. OSS/BSS transformation was also discussed – the key challenge is of course being multi-vendor and hybrid deployments, as most of the islands today are single-vendor and virtual only. To really tackle this challenge, providers are looking at their existing multi-vendor strategies and enhancing them to be able to cope with a larger number of vendors, as well as on boarding new ones faster.

How is TM Forum supporting the industry? 

TM Forum’s work in this area has been mostly dominated by planning for its transformation and assessing the impacts, up until the latter part of last year. That means we now have a wide range of assets, including an online training course, that can be used to get alignment across businesses on what is going to change and how. 

We are now deeply into the creation of the key assets that will be important in the deployment and implementation phase. We are currently building an end-to-end orchestration architecture with input from key CSPs and cloud service providers, we are working with other SDOs to agree some industry standard information models, as well as looking at areas such as process mapping, procurement and VNF packaging.