The migration away from static, hardware-defined networking to a more programmable, virtual and automated future is fragmented. The world’s network operators are all moving at their own pace and with their own priorities, while vendors are anything but unanimous in their vision of what SDN should look like.
It is not even clear what, at this stage, is driving the migration. Is it the need to slash expenditure on expensive hardware? Or is it in hope of the enablement of a new generation of profit-boosting services? Dimitris Mavrakis, principal analyst with advisory firm Ovum, is in no doubt that at this stage it is primarily the former.
“For the moment, I see SDN implementation as a pure cost-cutting exercise,” he asserts. “In an indirect sense it can be used to give more control to the enterprise customer, say by enabling the set-up of a software-enabled VPN in a more granular way. But there’s no evidence yet that it’s a major revenue stream for anyone. It will enable new value-added services in the future, maybe in areas like IPX. I think we’ll see adoption speed up once we see that revenue can be made.”
With SDN techniques, service providers can certainly expect greater efficiency through, for example, automated real-time traffic engineering. This should deliver increased network utilisation and ultimately translate to capex savings, believes David Noguer Bau, senior manager for solutions marketing EMEA with Juniper Networks: “But these capex savings won’t necessarily come by centralising all of the control planes, as the network re-architecting costs would be higher than the savings achieved,” he warns.
Picking the right moment
Chris Emmons, director of network planning with Verizon believes that timing is all: “If you are moving now, there’s a capex bill to pay,” he says. “But I’d expect, post-capex, for there to be cost savings. More dynamic provisioning of the network will mean less unused capacity.”
Operators currently spending to move in a software-defined direction for the sake of reducing costs will quite likely have one eye on other benefits emerging once the capex challenge is under control. There’s no doubt that as it stands, operators are too flat-footed when it comes to creating new products – certainly when compared to their agile OTT rivals. SDN will help in the long term to reduce this lag, helping traditional telcos prototype new ideas. It is not unreasonable to expect a shift from cost as a motivator towards the top line.
Not everybody visualises the revenue-generation dividend as a long-term gamble. Having been widely seen as a successful pioneer at the deployment phase of SDN, Asian carrier Pacnet claims it is now moving on to open new commercial inroads.
“With SDN we already do things faster, consistently across all our configs and with better utilisation of our network,” says Jon Vestal, Pacnet’s VP of product architecture. “The customer no longer has to buy based on peak usage. This is good for them, and for us as we have no dead air. It’s a win-win. As for the separation of the control plane and data plane, we’re split between serving old customers and gaining new ones. We’re offering what we couldn’t before. By breaking the wall between systems and application teams we’re winning new business.”
It could be argued that SDN needs to do better than just slash capex if it is to be more than a passing trend. Kelly Herrell, VP and GM of software networking with vendor Brocade, thinks that SDN – just as much as any other new technology – must stand or fall on its own commercial merits.
“I think at first operators were seeing SDN as a way to reduce reliance of human glueware and cut down on expensive errors,” he says. “Now it’s more about business benefits and the competitive angle. The cost-cutting imperative isn’t going to go away though. It’s just that nobody has ever been known to shrink to greatness.”
At this stage in its development, not even the most ardent of SDN’s backers are pushing for it as an all-or-nothing Year Zero upgrade, most being content to view it as a hybrid deployment. By the same token, many are adopting an SDN-like strategy in part of their network to both enable new services and reduce their capex investment.
Steve Bowker, VP of technology and strategy for TEOCO, a developer of software solutions for telecoms operators, is certain that SDN can be both things at once: “Once implemented, SDN will lead to far more efficient networking and hence significant cost savings,” he says. “It will also enable greater agility and new services.”
The decision on whether to implement SDN now or later depends on how recently an operator has made capex investments in its network, concludes Bowker: “Any operator currently investing and upgrading will be almost certainly starting to implement SDN, but those who’ve recently completed a major network expansion will most likely wait,” he believes.
The future’s probably bright
Even for those choosing to wait, SDN promises a list of operational benefits, from dynamic provisioning to guaranteed path diversity, improved load balancing and simpler network management. It also clears the way for much better visibility across a network, creating an opening for dynamic bandwidth-on-demand, bandwidth calendaring and other advanced high-availability service and traffic engineering possibilities.
But the route to this future is paved with uncertainty, not least owing to the robust internal dialogue currently raging within many carrier organisations. The wider debate about whether SDN is a cost-parer or business generator is being held in microcosm within many boardrooms.
“Most operators are not yet full SDN-ophiles,” believes Ben Parker, principal technologist at consulting firm Guavus. “There’s a CTO [a technology architect with an engineering focus] and CIO [responsible for the alignment of IT and business] separation here that overall is creating corporate uncertainty. It’s a gap that will close over time. You’ll only have to start seeing some real operational efficiencies and that will start to happen. Business opportunity presents itself and suddenly inertia moves out of the way.”
If there is disagreement within carrier boardrooms about what SDN really means, then at least there are signs of growing harmony across the industry, argues Paolo Campoli, CTO of service providers for EMEA with vendor Cisco.
“I see greater alignment on what’s driving SDN among service providers, and indeed vendors, than there was,” he says. “Concerns over areas like interoperability are diminishing. They’re now asking ‘Where in my business does it most make sense?’”
As we look ahead to the deployments of tomorrow we can probably expect SDN to remain to some extent a cost-saving weapon. SDN in its most conventional sense is a way to separate the control plane from the data plane, opening up network infrastructure to incorporate lower-cost hardware. What will change as the industry matures is greater move from SDN as a saver of capex to a saver of opex. And we can anticipate that this greater efficiency will have an impact on revenues.
SDN promises not only faster development schedules and speed of deployment, enabling a shorter time to market, it also offers a programmable ecosystem for all the applications that use the network. In order to exploit the full potential of SDN as an enabling platform and an ecosystem for applications programming, it will no doubt be necessary to improve the standardisation of APIs.
Transitioning to SDN will not be an overnight shift. Carriers are burdened with legacy hardware infrastructure, which is putting a brake on faster migration, meaning many will probably proceed with hybrid deployments. The full return on SDN investment cannot be expected for years to come, in most cases.
SDN will, paradoxically, reward both the bold and the patient.
