Virtual CPE: Bringing Pay-TV into the Internet Age

Murali Nemani

In the horse race to roll out advanced services, pay-TV providers often start with a huge handicap: The 10-year lifecycle for design, development and deployment of customer premises equipment that slows cable and IPTV while online video providers sprint toward the finish line like California Chrome at Churchill Downs.

The problem is one that’s familiar to anyone who’s ever made a bet at Belmont or Bay Meadows: No matter how smart you are, it’s impossible to predict what will happen when the race is underway. All too often the technology that’s the odds-on favorite at the start can turn out to be less of a sure thing over the long haul.

 Technology Cycle
In just five years alone, we’ve seen massive shifts in:

  • CPE horsepower—Proprietary silicon has given way to Intel and then to Broadcom architectures; support is required for escalating demands for higher-end graphics; and the 500 MB of memory that once was more than enough has faded in the stretch as applications routinely require 1 and now 2 Gigs of RAM.
  • App Frameworks—Within a five-year time frame, the Java application development framework has given way first to Flash, now to HTML5—and perhaps soon to native app development environments.
  • Browser Technology—A split in the browser community has led to the forking of Webkit thus fragmenting the browser world.

A solution that’s gaining traction in the enterprise community is to shed reliance on CPE resources and extend the concept of “Network Functions Virtualization” (NFV) all the way to the last node in the network – the customer premise. As noted in a recent Heavy Reading Insider report, this virtualization of CPE can reduce capex, accelerate time-to-market and service velocity, and enable service providers to dynamically update applications and services in the cloud.

Cablevision’s deployment of “cloud-based UIs for interactive TV services delivery” is cited as a watershed moment in CPE virtualization; Ziggo, CharterLiberty Global and Deutsche Telekom’s T-Labs and others have built on this early work by moving large portions of CPE functionality to the cloud. Yet other operators, even after virtualizing routers, transport and data centers, continue to place wagers on high-end CPE devices with heavy, PC-like cost structures that often are unable to keep pace with the changes in the industry.

Here’s our thought: By extending NFV to the customer premise and virtualizing CPE functions into the cloud, service providers can deliver user experiences that can keep pace with silicon innovation, browser updates and changes in authoring environments at Web scale—without the cost and time-to-market of replacing all of their existing STBs.

Although CPE virtualization has only begun lighthouse deployments, the Heavy Reading Insider report notes that it is “high on the agenda of network technology and device vendors.” Further, the capex and opex benefits are so compelling that “once they have experimented, operators will switch quite quickly to new operational models.”

For service providers who are fending off online competitors, we think it’s a bet they can’t afford not to make.