Virtual CPE (or vCPE) is one of the most popular use cased for NFV by service providers. It’s not surprising when you consider that the customer premise location, whether a business or a home, is the most complex and costly aspect for a service provider to manage. If you look at the situation today, often service providers have to deploy multiple discrete devices or appliances to deliver a service. And the range of services are pretty diverse, varying by region, market, vertical and customer. Services span everything including managed router, firewall, security (such as intrusion and malware detection), unified communications, application performance management….you get the idea.
In some cases, these capabilities may all reside as virtual appliances or agents in more costly and complex converged multi-service devices like integrated service routers. In any case, service providers have to deal with a lot of complexity. Challenges include navigating a myriad of business process and operational complexities spanning from order management to installation, administration and maintenance. Operational issues range from managing equipment obsolescence to service level agreement monitoring and troubleshooting. Identifying the root causes and remotely re mediating issues can be time consuming and costly. And with the uptake in cloud services there are more business critical (and performance-sensitive) application transactions traversing networks.
The outcomes are as expected: high operational costs, longer times to revenue, and ongoing operational complexity. All impact service profitability. Not surprisingly, virtualizing customer premise equipment (vCPE) has been a key target use case for NFV. A target market is small-to-medium enterprises, which collectively represent a very lucrative segment for service providers, with high recurring revenues. But all this obviously comes with some challenges.
By moving functions (like firewalls, application accelerators, and even routing, from the premise and into a data center) there is now even more critical transactions traversing networks. Incorrect QoS parameter settings, for instance, anywhere along the service chain could cause excessive latency causing some applications, such as interactive video, to experience unacceptable quality. Ensuring network configuration integrity end-to-end is even more crucial.
Partially for these reasons, we’re also seeing different types of virtualization strategies emerge. Virtualization-capable platforms are migrating outside the confines of the data center, all the way to the customer premise. This newer breed of device can host virtualized functions. Some of these functions are maintained and owned by the service provider, or perhaps the service provider is hosting these for the enterprise or another 3rd party. This multi-tenancy scenario requires that security integrity be preserved, and strong identity access management policies be in place.
Implementations will vary and there will be no single universal vCPE recipe. One thing is for certain, this is a rapidly developing space. We’re only scratching the surface of some of the operational challenges. These will be some of the topics we’re discussing at an April 30th roundtable discussion hosted by VanillaPlus. We’ll be joined with colleagues from Ericsson, Citrix, RedKnee, Comptel and Analysys Mason. Make sure to register.