Resident expert on switched digital video (SDV), Bruce Bradley, gave a briefing last week to talk about the bandwidth implications of SDV, deployment timelines, and how the technology is likely to evolve. Here are my takeaways from the conversation:
On the bandwidth front, cost models are showing that SDV is the most efficient (read cheap) way to free up bandwidth in the 2007-2008 period. No wonder cable operators are so enthusiastic, and willing to find solutions to issues with one-way CableCARDs in retail devices. Operators can reclaim a projected 50% of bandwidth with the technology. That leaves a lot more room for more HDTV.
As far as deployment goes, 2008 is expected to be the big year. (No surprise there.) But, we’re talking about North America, not globally. By the time SDV starts rolling out in Asia and Europe, we may be into the next phase of the technology. Specifically switched unicast, which is designed to provide personalized video streams (and new revenue opportunities via targeted advertising). Switched unicast becomes feasible when the number of set-tops in a service group approaches a 1:1 ratio with the number of video streams being served. Today a lot of service groups contain around 1,000 users. To get closer to the 1:1 ratio, that number would have to be more like 250.
If you’re looking at a timeline of technologies, the order of events will probably go something like this: switched digital video, followed by SDV with a variable bit rate, then MPEG-4 (less immediately urgent for cable operators now), then switched unicast, and finally cable GPON deployments. Not everyone will follow the same timeline in every region, but SDV is the logical first step.