Cloud Computing Will Empower EBIF
While EBIF user agents are allowing a base of legacy set-top boxes to handle interactive applications, they still are limited in terms of memory and processing capability.
Combining cloud computing with EBIF could be a way for cable operators to offer a common platform across all boxes no matter the age of the equipment, said Jeremy Edmonds, director of product management, ActiveVideo Networks, during an EBIF session at the Cable Show last week.
“Revolutionary technology, built, deployed, tested (and) enhanced continually in the Web world, is now accessible in set-top boxes. The amount of innovation able to be done in the cable environment is increased by orders of magnitude,” said Edmonds, the sole non-Comcast panel member.
The concept is similar to Google where the applications are processed by powerful servers. In a cable scenario, the visual application would be rendered in the cloud with the output being a visual image sent to the set-top. The file would be around 13kb, for example.
“There are significant (EBIF) applications that don’t need cloud computing with RFIs (requests for information) and voting applications. This is not an answer for revolutionizing the industry,” Edmonds said. Rather it is a way to allow the entire EBIF footprint to run the extra features advertisers want to utilize.
Comcast’s three panelists talked about network and application challenges as well as lessons learned. For starters, one problem has been that individual applications treated as discrete instances all have overhead costs.
Mike McMahon, principal engineer, new media product development, Comcast Media Center, suggested the creation of an interactive television template based on a non-proprietary open standard. This would allow non-technical personnel to create derivative instances of an application. Being non-proprietary, all vendors could use the same customization tool to provide templates.
Otherwise, proprietary templates mean more purchases and asking the app developer to make the changes also adds cost and other problems. “Is it really the same code base as the first one? Did the guy make a typo? Did he change it….You’ve got to treat it as a new app. You’ve got to test it,” McMahon said.
Canoe Ventures is actually working with templates. (For more, click here).
As for the technical side of EBIF, there are specific challenges with the data signaling path, for example. “Legacy signaling path devices out there are scarce resources and need to be treated as such,” said Kevin Taylor, fellow, Comcast Cable.
With EBIF applications there is some concern over synchronized events, where a whole host of set-top boxes respond, something that operators have not had to deal with since services like VOD tend to have more randomized responses throughout the day. “EBIF creates a toll for synchronization that we need to manage,” Taylor said.
These applications could be de-synchronized, said Robert Dandrea, principal engineer, Comcast Cable. “If you vote, (there could be) a random backlog. Your message will leave (the set-top) within several minutes. This needs to be built into vote tallies.”
Comcast also learned there is a need for an OSS system to “snoop” EBIF both at the Comcast Media Center and local muxes, monitoring whether the messages are being sent and received. “The user agent (also needs) the ability to receive queries and reply with details about what is going on…(You) can query a set-top on demand and find out what it is doing,” Dandrea said.
Another “major” change Comcast made is to require all of its markets to receive EBIF national programming from the CMC rather than directly from the programmer. The idea was to minimize satellite infrastructure and MPEG2 encoding gear costs, Dandrea explained.
Comcast, which began deploying EBIF apps a couple of years ago, plans to launch several new bound apps this year, including some with Canoe. Among them are guide enhancements that will allow them to add features without a monolithic release and some widget-type applications.