The telecom industry is several years into its own journey, transforming network infrastructure using concepts from Network Functions Virtualization (NFV) and Software Defined Networking (SDN). However, early adopters are still ironing out the kinks, and in the process, are having periods where it seems like they undertook a far harder task than realized. This is to be expected, because building complex networks with a new set of hardware and software components, while ensuring interoperability that satisfies the industry’s expectations is a totally new paradigm.
Our upward climb is similar to the IT world’s experience with virtualization in the late 1990’s. When they began, there were all kinds of performance and compatibility issues, and disagreements over which virtualization technology to use (e.g., OS virtualization, hardware emulation, and paravirtualization). Turn the page 15 years forward, and now virtualization is everywhere. A technology that started as a way to get better server utilization is now essential to data center management and automation.
New Service Models Needed
In our industry, communications service providers (CommSPs) are seeing data traffic far outpacing revenue, and the situation could get worse as consumers watch more video on smart phones. Much video traffic originates from the Internet, and global video traffic is forecasted to be up to 80 percent of all consumer Internet traffic by 2019, according to a Cisco forecast.1 With data plan profitability falling (measured by price per GB), CommSPs need to offer new revenue-generating services, and network transformation can help in various ways, including:
- Instant messages and email services: Give customers what they want in the way of enhanced messaging services. One example is Libon, a mobile service developed by Orange Vallée (a subsidiary of Orange S.A.).2
- Deep packet inspection: Offer data security services that many organizations and corporations are willing to pay for to catch malware before it has a chance to infiltrate their internal networks.
- Wi-Fi service on trains: Help end users stay connected by supporting Wi-Fi service on trains, where an Internet service provider runs over-the-top (OTT) of CommSP cellular networks for a fee. NFV and SDN enable CommSPs to more easily set up this type of service in network access points.
As a CommSP, network transformation provides the agility you need to optimize your network resources. You’ll be able to automate network function deployment and termination on-the-fly, wherever these functions run: in the Telco Cloud, in the data center, or at the edge.
The development of such capabilities is making good progress. Along these lines, at February’s Mobile World Congress, around two dozen service providers and solution vendors announced their intent to join the Open Source MANO (OSM) Community. The focus is on delivering an open source Management and Orchestration (MANO) stack, aligned with the European Telecommunications Standards Institute’s (ETSI) NFV Information Models.
The industry saw a demonstration of the OSM “project ignition” that implemented seed code contributions, used commercial VNFs, and proved the key guiding architectural tenets of OSM such as layering, abstraction, modularity, and simplicity. It also delivered functionality for end-to-end automation, SDN underlay control, multi-site capability, multi-VIM capability (OpenStack and OpenVIM), and enhanced platform awareness (EPA).3
To the surprise of some, open source software isn’t free source code, so to make it work as needed, CommSPs may need support from the open ecosystem. And for those who ask if the Telco Cloud is ready for prime time, it’s important to remember it requires operational re-structuring and a mindset change, or in other words, we are on an evolutionary path.
Choose the Right Hardware
One misnomer is that in a software-defined world, hardware no longer matters. But the truth is agility comes with hardware ingredients and optimizations, and EPA enables the discovery of these features, allowing network functions to be deployed on the appropriate server (i.e., matching workload requirements to specific server capabilities). Time-sensitive applications may need to run on servers specially designed with features that minimize latency. One example is Intel® Resource Director Technology available on select Intel® Xeon® processors. The technology gives software programmers fine-grain control over the processor cache to prevent time-critical data from being evicted by other processes, thus providing more deterministic performance.
The Future Is Bright!
When we’re all working hard to get solutions ready for field deployment, it’s helpful to remember the end goal for our journey – delivering future network capabilities that provide the agility customers demand, with the ability to rapidly deploy value-added services that generate new revenue streams. As future 5G network capabilities roll out and new devices require connectivity, as in the case of IoT – the groundwork of today establishes the most optimal network possible for harvesting the future opportunities.
Learn more about accelerating network transformation here.
You can follow Lynn on Twitter at @comp_lynn.
1“Cisco Visual Networking Index: Forecast and Methodology, 2014-2019 White Paper,” 2015, pg 2.
2GIGAOM website, “Orange’s Libon messaging app lets you chat with people who don’t use Libon.”
3Telefónica website, “Telefónica joins 22 service providers and solution vendors to create global Open Source MANO (OSM) Community,” February 22, 2016.