Virtualization: Not if but when
Mark Grayson on how the Forum is addressing virtualization
You can watch Mark Grayson’s webinar on Network Function Virtualization here.
Virtualization – and in particular network functions virtualization (NFV) – is in some ways a relatively new subject: ETSI, for example, only began formal discussions of NFV in late 2012. However, it has been clear from an early stage that virtualization is a transformative technology, changing the way network infrastructure is deployed in numerous contexts.
That is why, since last year the Forum has been discussing virtualization and how it affects and enhances small cell propositions for, let’s be clear, virtualization of small cells is not a question of if but when. This is especially important in the rural and remote context where NFV can help operators to offer scaled-down deployments that are still economically viable.
The point our work has so far reached is the subject of a webinar I presented at a recent MENA operator meeting, as well as a Release Five paper on Virtualization in Small Cell Networks [SCF154] but there will be further updates in due course leading to an important publication in June in time for the Small Cells World Summit.
The main jumping off point for this sort of discussion is a question: small cells need to interwork with the macro layer, which is increasingly being virtualized, so is there a common virtualization approach – a common architecture that can span the complete HetNet?
It’s a simple question. And one of the simple responses is that we can’t virtualize all of the small cell, or the complete macro cell for that matter. But this is where things get slightly more complicated. We need to decompose a set of functions that today exists within a physical small cell unit and split those functions between a virtual network function and a physical network function. How we do this is already one of the key study items of Small Cell Forum.
It’s also the question my webinar addresses. It takes in related areas like radio performance, transport delay, bandwidth expansion, LTE-advanced features, VNF hosting, the overall end-to-end system architecture able to support virtualized small cells and the requirements of 5G. We also need to understand how and whether virtualization can support innovation across a multi-vendor ecosystem. We are already seeing novel decompositions from a broad set of small cell vendors, all currently based on proprietary approaches. A key question we have to ask is whether we can leverage the Forum’s success in driving multi-vendor ecosystems to ensure we don’t lose that aspect as we transition into a virtualized world?
This is, you’ll have noticed, a complex issue. Certainly, compared to Releases One to Five a study into virtualization involves a new way of working for Small Cell Forum. However, it also requires a set of competencies that we think the Forum is well suited to offer – as SCF154 has already indicated.
For the moment we’re not discounting any particular split of functions. We’re looking at many different options and the linkages between virtualization, the business case, transport characteristics, deployment-related issues, and the critical multi-vendor aspect. The plan, as I have said, is to be able to publish a major study in June this year.
In the meantime please look at our work to date, and contribute if you can – or ask questions. I expect to be asked quite a few at the Singapore Plenary and the Small Cell World Summit – and I’m looking forward to answering them. But however you choose to take part in bringing together small cells and virtualization, please do get involved. The rewards for getting this right are too great for any of us to take them for granted.