The Rise of Virtual EPC
Any customer of Heavy Reading research services will know that network functions virtualization (NFV) is now one the most critical initiatives in telecom networking. The technology applies to virtually all parts of telco network and the Evolved Packet Core (EPC), in particular, has emerged as one of the more tangible examples of how NFV can be useful -- even transformative -- in carrier networks.
Virtual EPC was an emergent theme at the 2013 Mobile World Congress, with Cisco, NEC, and Intel showing demos, and at the 2014 event it was one of the biggest networking stories at the show. AT&T's announcement that it would work with virtual EPC startup Affirmed Networks epitomized this. (See AT&T's Cloud Future Takes Shape.)
There are, I think, two primary reasons for this interest in virtual EPC: (1) because EPC is a substantial enough application that it would present meaningful progress for an NFV initiative; and (2) because the nature of the application lends itself to virtualization, making virtual EPC "doable" as well as meaningful.
In my new Heavy Reading report, "The Rise of Virtual EPC: A Mobile Packet Core Forecast & Analysis," I examine these drivers in more detail.
We know from our mobile operator research that demand for virtualized mobile core is real. There is enthusiasm from mobile network operators for a software-centric core network that is programmable and can be more easily customized and optimized for end-user needs. In principle, they say, the model is very attractive.
Equally, it is clear that there will be a long period of development of virtual EPC, for reasons of technical maturity and operator investment cycles. In practice this means adoption will be piecemeal and initially focused on less-critical, smaller-scale use cases.
In the first instance, virtual EPC will tend to be deployed in parallel to the main production EPC and will focus on machine-to-machine (M2M) and customized enterprise services. These types of services can have specific traffic profiles that might benefit from a particular EPC configuration (and specifically, a virtual P-GW configuration), and they provide an opportunity for operators to experiment with a virtual core without putting mass-market services at risk.
There is also a substantial near- to medium-term opportunity for virtualization in what is sometimes called the SGi-LAN Ė the services complex between the 3GPP-defined mobile core and external networks. These functions, such as DPI, Web proxy, video optimization, content and charging gateways, and firewalls, can be virtualized and mapped to software-configured service chains to create a more efficient and programmable value-added services domain. Startup ConteXtream is one of the pioneers of this concept.
Over time, there's potential for both approaches to evolve and, eventually, to transform the mobile core. For example, it will be possible to combine SGi and EPC functions into bespoke, software-configured service chains. And virtualized P-GWs, initially targeted at early adopter enterprises, will scale to more customer types and be used to add capacity to (and in time start to replace) the "classic EPC" as the technology matures.
In terms of vendor positioning, all of them now endorse the idea of virtual EPC, and over the past 12 months, spurred on by operators, most have come to embrace the opportunity wholeheartedly.
I'll write more about this in a follow-up column, but for more detail, the full report covers the following vendors: Affirmed Networks Inc., Alcatel-Lucent, Cisco Systems Inc., Connectem Inc., Ericsson AB, Huawei Technologies Co. Ltd., Mavenir Systems Inc., NEC Corp., Nokia Solutions and Networks (NSN), Samsung Group, and ZTE Corp.
ó Gabriel Brown, Senior Analyst, Heavy Reading