The concept of network virtualization is one of the hottest topics in the telecoms sector for a variety of reasons - primarily because it’s the most radical innovation to happen to the telecoms network in decades. As networks evolve into all-IP environments, just about everything will eventually be software-based and hosted in the cloud - to include network functionality and control.
That’s where NFV comes in. The basic idea is to take functions usually embedded in network hardware and turn them into software that can run anywhere - centrally controlled, orchestrated, and managed. The benefits are well documented now: network virtualization enables point-and-click installation (which enables greater business agility), and allows telcos to expand their networks as needed. It also saves a bundle on capex/opex and helps telcos avoid vendor lock-in (some vendors still push for a single-vendor approach, but the point is that operators now have the option of choice).
There are numerous applications for NFV, but one of the more striking apps is for the interconnect market. Which brings us to session border controllers (SBCs), which - like the most essential networking technologies in the telecoms space - have a reputation for not being all that sexy as tech topics go. But combine SBCs with NFV, and you have a recipe for turbocharging the interconnect space with virtualized points of presence (PoPs).
Telecom Asia e-Brief: NFV
Like the rest of the telecoms sector, the interconnect market is going through some heavy changes thanks to the massive growth of data traffic and the accompanying economic changes. More and more carriers of all sizes need to interconnect with other operators to extend their reach, but interconnect carriers are seeing traffic grow but the cost per voice-minute declining.
Meanwhile, the service landscape itself is changing, and offering simple voice, SMS, or data services is no longer enough. Telcos increasingly seek to offer a wider variety of new digital services - from OTT content and managed enterprise services to e-commerce, healthcare, and the Internet of Things - at varying levels of capacity requirements and service quality. And while voice traffic isn’t the money-spinner it used to be, it’s still the service that’s most in demand, even in IP form.
Interconnect carriers must be able to support this new agile service environment by becoming agile themselves, optimizing their networks and reducing costs. And one way to do that is by leveraging NFV to create virtualized PoPs.
A “virtualized PoP”, in its simplest form, is a software-based SBC. SBCs are designed to regulate and transcode VoIP/media traffic. And as the name implies, they’re usually deployed at the borders of networks, whether that’s the internal border between access and backbone networks, or between different service provider networks. But like most network appliances, they’re hardware-based, expensive and inflexible. And they’re incredibly complex.
That’s why virtualizing an SBC is no mean feat - the software must support signaling functionality and data plane functionality on any hypervisor on top of bare metal, as well as reliable scheduling, resource allocation, and fast packet I/O for media handling and signaling DDoS protection. And it has to do all that in a virtualized environment rather than on proprietary hardware.
Of course it’s possible - there are already carrier-class virtual SBCs on the market. And virtual SBCs can be deployed as a virtual network function (VNF) for interconnect purposes. Put simply: what used to be a rigid, pricey box is now software that can be deployed on cheap commodity servers in a data center rack. An NFV-enabled virtual SBC can serve as an IP interconnect point (that is, a virtualized PoP) between the service provider’s customers and its backbone network.
Result: an interconnect PoP that’s cheaper to install and can be set up in a matter of minutes from a desktop. (Compare that to the up-to-18 months it takes to install the hardware version.) Which is exactly the kind of flexibility carriers need in an agile service environment.
All up, the rise of virtualized PoPs promises operators lower start-up costs, faster deployment, and new markets to tap into. It also allows them to deploy closer to their customers’ customers.