A practical and pragmatic look at the development and final roadblocks for Software Defined Networking and Network Function Virtualization Technologies, with Brett Brock
With AT&T, Verizon, CenturyLink and NTT filling up the column inches with SDN and NFV announcements, the future certainly looks rosy for a potentially industry defining technology
But for every step forward, we don’t seem to be getting any closer to live deployment, and an industry wide movement towards virtual network delivery. So what is actually holding us back?
“Economics” commented Network Design Engineer Brett Brock
“Delivery on the Communication Service Provider expectations of huge cost saving will shrink vendor revenue” he added. With Telecommunications revenues flattening, any technology which can reduce costs on either (or both) CAPEX and OPEX has been with great excitement
“Whilst this has been accepted as the way forward, the introduction of SDN & NFV will also constrain the vendors’ ability to invest in reinventing themselves as software first enterprises – it’s a surprising conundrum for the vendors, which could have had a stuttering effect on the growth of the technology” Brett highlighted
But this in itself raises an interesting question – what are the main drivers behind the development of SDN and NFV?
Firstly the main benefits were touted as a reduction of CAPEX and OPEX, but industry commentators have questioned whether this is enough of a justification to completely redefine the way we design, build and manage networks. Before we see a major transition through to virtualized and automated networks, most believe we need to get a better understanding of how we can draw out the other benefits of both technologies (scalability, flexibility etc.).
Until these questions have been answered, is this simply a case of Evolution for the sake of Evolution?
“While I don’t see SDN and NFV as “Evolution for the sake of evolution”, ETSI PoC’s have been geared towards operational efficiencies and execution speed” Brett highlighted to us “NFV is a matter of controlling one’s own destiny”
“The economics of the current carrier technology and technology investments are unsustainable in the face of flattening revenues. NFV offer vehicles for preserving margins by driving down cost long-term, but it requires rather high an initial investment. Furthermore, it will not fill product pipelines with no or better ideas”
As Brett highlighted, there are certainly short and long term benefits to the technology, however the flexibility (which SDN & NFV promise), scalability (which SDN & NFV promise) and the cost reductions (which SDN & NFV promise) can only truly be delivered through Opensource
A slightly more sensitive question to ask would be how committed the vendors are to moving forward with these initiatives. Although Opensource is the spirit and benefit of both SDN & NFV, this is another aspect which would constrain the vendors’ profit margins (to a degree)
“To commit to open source projects vendors have to commit to disrupting their own businesses. Instead of competing on established rules they helped shaped and have mastered, they would be forced to compete on terms defined by OTT players, developer-led start-ups, and the web titans” Brett told us “Economically speaking, the dominant telecom vendors’ business model’s requires considerable more revenue per unit sold to keep them in the black”
Recent reports in the press suggest that some of the vendors’ are not as committed as others – Juniper and VMware left the OpenDaylight project in June, though Brett isn’t convinced that the vendors are entirely to blame
“Vendor lock-in is a matter of economics and human nature; it is just easier to assign blame to vendors. Job posts ask for years of experience with specific tools. Employers pay “subject matter experts” a premium for rare depth of knowledge on a given platform. The behaviour and more set extremely high exit barriers for swapping out a foundational platform, open source or not”
Brett Brock will also be speaking at IIR’s 3rd Annual Network Virtualization Forum, North America, which will be taking place 6-7 October in Dallas, to discuss all of the issues raised here today
To have a look at what Cow Communications, Equinix, AT&T, Verizon, NTT North America, SingTel and Telus are discussing at the forum, you can download the full e-brochure right here