Video Distribution: Maybe One Size Doesn’t Fit All?

An examination of Cisco’s latest annual internet use and forecast report and a recent Streaming Media article highlight a problem that is mostly acknowledged only by engineers whose job it is to manage internet-delivered video transmissions.

The problem? Insufficient bandwidth for a voracious consumer appetite for video and other data intensive programs.

It’s not that broadband speeds aren’t increasing. They are. The Cisco report forecasts that broadband speeds will nearly double by 2022 over 2017’s worldwide average of 39 Mbps. So why worry about internet capacity for the ever-growing video traffic increasingly viewed on stationary and mobile devices?

Because, to date, streaming services only employ inefficient unicast delivery for stored and live content, and that method is straining network capacity in many parts of the world. DTC has written about this for some time, and some streaming-community acknowledgement that hybrid broadband/broadcast technologies should be part of the TV ecosystem is a welcome development.

This is especially true for streaming of live content. With few exceptions, live streaming only occurs in SD or HD, leaving 4K UHD sports and other live events mostly out of the picture. In the parlance of the streaming community, unicast doesn’t “scale.”  A delivery mechanism that “scales,” broadcast or multicast, has been hiding in plain sight. However, current production and transmission technologies can’t move IP-ready content over the broadcast airways, but that has changed with long-overdue R&D to create new terrestrial systems.

One of those systems, ATSC 3.0, enables the broadcast of IP-compatible content. There is a growing recognition that satisfying consumer demand for high-quality video with hybrid broadcast/broadband delivery could be a common-sense solution where unicast alone can’t provide it.

For those who wonder if next-generation unicast technologies will solve the capacity/quality problems, consider these data points from the Cisco report:

  • Globally, IP video traffic in 2022 is forecasted to be 82 percent of all IP traffic compared to 71 percent in 2017.
  • A “cord cut” household in 2017 used, on average, 141 GB per month, while an average household used 82 GB per month. The greater inefficiency for all-streaming households, according to the Cisco report, is “because linear television generates much less traffic that the internet that is unicast to each video device.”
  • By 2022, Cisco forecasts that globally, 63% of households with broadband will have 50 Mbps service. But, consider the bandwidth demands cited by the report for today’s and for future applications:
    • 51 Mbps for IP UHD
    • 100 Mbps for 8K “wall” TV
    • 167 Mbps for HD virtual reality
    • 500 Mbps for UHD virtual reality

It’s not a stretch to conclude that many future data-intensive applications will create intractable traffic jams for unicast networks. Even with significant improvements in unicast networks, it’s difficult to see how one-to-one transmissions can handle all content types to massive numbers of devices.

Displays will only get larger and accommodate greater lines of video resolution, immersive audio and other data-intensive tricks that have yet to be created. It seems foolish to dismiss out of hand more efficient multicast distribution. And now that a new over-the-air TV system can broadcast IP-compatible content, the streaming community might want to reconsider a future that continues to employ unicast-only distribution.