The five-part flow of an interactive application

We've now created a "bound" EBIF app—specifically, an RFI (request for information). The RFI was scheduled into the participating network's traffic system, and injected into the video bit stream using a data carousel. Next, that application enters the set-top box. What happens next is in the domain of the User Agent, or UA—a small kernel of software, designed to "see" and render the clickable object, on the TV screen. There are multiple providers of UAs, from a growing vendor community.

Someone clicks. The RFI was rendered on the TV screen, and a viewer clicked. She wants more information. Then what? First, the viewer is prompted to re-affirm interest—which, in so doing, serves as an "opt-in" acknowledgement.

Then, the information contained in that click is returned to an aggregation server at the MSO headend, using an out-of-band signaling path over the cable from the set-top box.

From there, the fulfillment process can begin—in this case, implementing the sending of the information requested. If the EBIF app had been an advertisement, for instance, or a clickalong with a TV show, the information representing that activity would also travel out of band to an aggregation server, for use in analytics and to develop different versions.