 | Issue: | EMEA 2014 |
| Article no.: | 11 | |
| Topic: | NFV and SDN: What happens next? | |
| Author: | Martin Taylor | |
| Title: | CTO | |
| Organisation: | Metaswitch Networks | |
| PDF size: | 214KB |
About author
Martin Taylor, CTO, Metaswitch Networks
Martin Taylor has spent over 20 years in the telecom and network equipment industries, with diverse experience in product marketing, engineering, technology planning and business development. At Metaswitch, he has led product strategy in hosted business services, web services, APIs, UIs, software-based SBC and unified messaging, plus the evolution of the product line for both IMS and pre-IMS networks. Martin is driving high-level strategy towards NFV and SDN with near-term deliverables of the industry’s first virtualizable carrier-class SBC and the industry’s first implementation of IMS core functions in a cloud environment.
Martin’s prior experience includes the role of founding CTO at CopperCom, where he was widely recognized as an early pioneer in Voice Over Broadband for his leadership of the ATM Forum standards work on Loop Emulation.
Martin is a graduate of the University of Cambridge, England with a degree in Engineering.
Article abstract
Network Functions Virtualization (NFV) and Software-Defined Networking (SDN) are probably the two hottest topics for network operators right now. These technologies promise new ways of building communications networks at lower cost and with greater scope for innovation in network services.
This vision of the future will only be achieved by network operators that embrace NFV and SDN fully and are prepared to implement fundamental organizational transformation to make the most of what they enable.
Full Article
The rapid evolution of standardized commodity computing infrastructure over the last few years has created an opportunity to build communications networks in totally new ways.
In the past, the telecoms industry built on special-purpose network elements based on proprietary hardware. This will give way to a future in which network functions are implemented almost entirely in software running on shared pools of standard hardware resources, just like enterprise, cloud-based IT workloads. This virtualization of network functions will enable new services to be created, deployed and managed far more quickly and at far lower cost than has been possible in the past. The programmability of the virtualized networking environment will go hand-in-hand with a new degree of programmability of the network functions themselves, enabled in part by a clean separation between control plane and data plane functions.
This software-centric future is exciting and Network Functions Virtualization (NFV) and Software-Defined Networking (SDN) are probably the two hottest topics for network operators right now. These technologies promise new ways of building communications networks at lower cost and with greater scope for innovation in network services.
Regardless of what kind of network you operate – fixed, mobile, cable, metro Ethernet, long-distance interconnect – or what kinds of services you offer – data, voice, messaging, content delivery, virtual private networks – NFV and SDN offer opportunities to transform the economics of your network while at the same time accelerating your ability to design and deploy new service capabilities. NFV and SDN represent the most significant pivot in the telecom industry since the transition from TDM to packet got under way a decade or more ago.
There are a lot of topics related to NFV and SDN, but the focus of this article is on the benefits of software-based network functions and a look ahead to what impact NFV and SDN will likely have in the next five years.
What are the benefits of software-based network functions?
Software development is not nearly as capital-intensive as hardware development. Furthermore, software lends itself to a process of incremental improvement that is simply not possible with hardware. As a result, we can expect a much wider range of vendors to participate actively in the market for software-based network functions in comparison with traditional network boxes. Competition among these vendors will result in a healthy marketplace, offering network operators a far greater range of choices than before. And the pace of innovation in network functions is bound to accelerate, propelled both by increased competition and by the shorter release cycles that are typical of software-only products.
Network operators should also find it easier and less costly to approve new software-based network functions for deployment in their networks, in comparison with the approval process to which new hardware needs to be subjected.
With software-based network functions, it finally becomes feasible for network operators to drive service innovation by conducting small-scale trials of new service concepts. The capital cost exposure of conducting such trials is minimal, since they can be run on spare capacity provided by existing hardware resources, and the risk of disruption to existing services is minimized since the trials can be conducted inside what is effectively a ‘sandbox’ that provides insulation from other established services.
An experimental service that proves unsuccessful can quickly be shelved without stranding any hardware investment, while providing valuable lessons in what works and what doesn’t. On the other hand, a new service concept that successfully gains market traction in early trials can be put into production and scaled up very quickly. The philosophy of prototyping new service concepts with the mantra of “fail fast, fail early and fail often” has been a key factor in the success of many over-the-top services, and software-based network functions enable network operators to embrace this philosophy for the first time.
All of these effects combined together should result in a dramatic improvement in the rate at which network operators can innovate in the service space. This is obviously hugely positive for the industry.
What impact will NFV and SDN have over the next five years?
The advent of NFV and SDN represents perhaps the single most important technology event in the telecom industry since the arrival of digital switching. Network operators that choose to embrace NFV and SDN fully have the opportunity to radically transform their businesses both to reduce their cost base, and to become far more agile in their ability to introduce new services. In short, to become ‘software telcos.’
In five years’ time, we can expect to see those network operators enjoying the benefits of NFV and SDN as follows:
● Almost all new network build-outs are based on highly commoditized hardware, combined with best-of-breed software – often sourced from specialist vendors. This will be true not just of network functions that can be deployed on standard x86 server hardware, but also of switching and routing elements (based on white box switch hardware) and even optical transport devices. Major reductions in network capex will result from this.
● Almost all new services are created by deploying new software elements in the network, or by combining existing software elements in new ways. It will be possible to prototype new services in a matter of days, and bring new services to market in weeks. It will be possible to respond to feedback about new services far more quickly than it is currently. New services that are successful can be scaled very rapidly by leveraging existing commoditized hardware, while new services that fail to gain traction can be stood down without the loss of major investments.
● Automation of service deployment reduces lead times for delivering services by orders of magnitude. Putting in place an enterprise VPN service for a given customer, for example, may take weeks. New business process applications are able to leverage the programmability of the network infrastructure using the new APIs exposed by it, shortening that provisioning time to hours, and even permitting a degree of customer self-service provisioning.
● The flexibility of a virtualized infrastructure enables network operators to offer entirely new types of services. For example, a network operator can address the needs of an enterprise customer that wants some specialized communications service by creating a virtualized service instance specifically for that customer, and customizing it accordingly. This can be accomplished far more quickly, easily and safely than attempting to satisfy specialized needs by making incremental software enhancements to some common, shared service platform based on proprietary hardware.
● Standardization of hardware and automation of operations management dramatically reduces operational costs. For example, in a conventional network, the failure of a piece of hardware is typically not service-affecting, but urgent action is required to replace the failed hardware in order to restore the proper level of fault tolerance. With NFV, the failure of a piece of hardware has no more impact than a temporary reduction in the maximum capacity of a given service, which is accommodated within planned hardware capacity headroom. There is no urgency to replace the failed hardware, and the procedure for doing so is extremely simple and completely standardized.
This vision of the future will only be achieved by network operators that embrace NFV and SDN fully and are prepared to implement fundamental organizational transformation to make the most of what they enable. This will not be easy: decades of procuring and building networks in the traditional way have created a culture inside most network operators that may put up strong resistance to the kinds of changes that are needed. But network operators that make the most of NFV and SDN will find themselves in a far better position than they are in today to compete with over-the-top services and with aggressive new entrants into the network business. We are at the dawn of an era of dramatic change in our industry. Never before have we seen such strong pressure to evolve, and to evolve rapidly. NFV and SDN are the tools we need to enable that evolution, and we ignore them at our peril.
