We should really start calling NFV the “Softwarization of Network Functions.” At the recent LTE Conference held in South Africa, I teamed up with Professor Thomas Magedanz of the Fraunhofer FOKUS, along with Dr. Marten Scheffer, GM of Network Engineering for Neotel and Vivek Bhargava with Oracle to discuss that very trend within network operators on a panel entitled, “NFV and SDN: Hype vs. Reality.” Thomas actually introduced to us the Google-worthy term “softwarization,” which I think is a very descriptive turn of a phrase relating to what’s going on in the industry, especially since 90% of the audience was really unfamiliar with what NFV actually was. And while there was lack of familiarity with NFV, virtualization, on the other hand, was definitely familiar to the crowd.
What was even more interesting was the fact that, unbeknownst to many in the audience, agile operators like Neotel were already embracing and implementing software defined network concepts in the African market within their infrastructure to make rolling out services faster and more efficient. Dr. Scheffer talked about how Neotel has already seen the benefits of software based infrastructure and networking, and it will be an integral component moving forward to help reduce cost, streamline operations and roll out services faster. How fast? Instead of taking hours to configure new applications, it’s now happening in seconds.
“What should operators expect from NFV and softwarization?”
Dr. Magedanz, who was chairing our panel, posed this question. Traditionally, there are defining events that compel service providers to make a technology change in their networks, and in most cases, it’s due to operational (OPEX) and CAPEX saving. But when it comes to NFV, the ability to become an agile operator is one of the key things driving them.
How does NFV make operators more agile?
Faster time to market or rather increased service velocity is one aspect. By implementing network functionality in software, you’ve eliminated several steps in the deployment, testing and trialing of new services and new functionality in the network. A good example is a WebRTC OTT service based on a new release of a media server that supports a new codec such as VP 9. New firmware, or worst case, new DSP boards and servers would first be installed if this had to be done in hardware. This hardware would undergo a series of tests in a lab environment with simulated traffic and then ultimately be moved out to limited production environments requiring more hardware installation. There would be limited trials performed and if all went well, the operator would install the balance of hardware platforms needed to support a wider implementation of the service.
Alternatively, in NFV environments where the applications are no longer tied to purpose-built hardware, the functionality can be spun up in a virtual machine, tested, and plied with a limited amount of live (albeit friendly) traffic to validate performance. We’re now talking weeks or days instead of months to deploy new revenue supporting functionality on what will effectively be the production virtualized environment. This is all done without the need to install dedicated hardware-specific capacity throughout the network. This approach substantially lowers the risk and cost of failure if the service doesn’t take off, enabling operators to roll out more services in a shorter amount of time to more focused verticals or demographics.
Operational agility is also enhanced. Upgrades to functionality can be applied in a more agile manner to the applications, and backed out very quickly if something does go wrong. In addition, with the focus that NFV puts on automated lifecycle management and service orchestration, applications can self-monitor to indicate when additional capacity is needed (or not needed) and then scale out or scale down resources without human intervention. This focus on automation helps eliminate the cost of stranded resources that’s currently encountered using physical, discrete network functions. We’ll be talking more about automation of lifecycle management in an upcoming SDx Central Demo Friday webinar – click here to register.
What do you expect from to get by moving to Network Functions Virtualization? Tweet us your comments using the hashtag #NFVAutomation at @Dialogic.