Adaptive Streaming: Cloud, Gateway or Both?

Publish date:

“Where does adaptive streaming happen
— in the cloud or in the gateway?”

Adaptive streaming is the process of
slicing a digital video file into different
sizes: maybe one that’s two seconds
in length, another that’s four seconds
and so on. The technique necessarily
works hand in hand with “transcoding”
and “transrating,” which match
available bandwidth with video stream size. Less
bandwidth, stream the smaller file; more bandwidth,
stream the bigger file.

One glance at last week’s TelcoTV event in New
Orleans shows a vendor community that is teeming
weth adaptive streaming (and teaming, for that matter
— see Itaas, RGB and Verimatrix). Why? When
one is bandwidth-constrained, which can happen
squishing a linear, multichannel HDTV product over
DSL networks, it’s good to have adaptive streaming
at hand.

Which isn’t to say that adaptive streaming is
just an AT&T thing. It’s attractive to anybody with a
network that’s staggering under the weight of that
pesky 45% growth rate of video as a percentage of
IP traffic. That means cable, mobile, and telco networks.

Which brings us back to you, Jan. Where does
adaptive streaming happen, gateway or cloud?
The answer is … yes. Turns out, parts of adaptive
streaming can happen in both places.

Some cable technologists say the headend is
the place to right-size streams to suit the care and
feeding requirements of the end displays — tablets,
PCs, laptops, things that connect through the cable
modem, not the set-top box. Slice, authenticate and
secure video streams there.

It’s all very Scotty McNealy, who put Sun Microsystems
on the map in the late 1980s with this famous
remark: “The network is the computer.” (Which
brings to mind the one about how the difference between
early and wrong is indistinguishable.)

Other technologists point out the gateway or “hybrid
box” (half cable modem, half set-top box) could
be the place to do transcoding. In that case, it’s not
so much about sizing the stream for the screen,
based on available bandwidth. It’s more about making
it so that any of your screens can play any of your
digital video archives.
Maybe some titles you own were compressed
with the older stuff of MPEG-2.

Maybe some of your
screens only know how to decompress the newer
stuff of MPEG-4/H.264, or vice versa. That’s when
you’d consider putting transcoding activities there.

Plus, silicon is simultaneously growing in capability
while shrinking in size. The gates on the chips
are there (or will be) to support the heavy processing
requirements of trans-coding.

Which is good, because there’s an H.265 compression
method on the drawing boards, and the
2012 Consumer Electronics Show is likely to show
fresh examples of “4K” video — the next size up in
picture resolution, from today’s 1080p HDTV.

Anyway: Gateway or cloud? Yes. Both. So far.

Stumped by gibberish? Visit Leslie Ellis at