momentum of IT and
cloud methods can’t be
are largely cloud/IP-based today. Examples
are Netflix, YouTube, Hulu, Watch ABC and
many others. However, media facilities—some production, post-production, broadcast ops—still have strong components of AV-specific gear. Much of this infrastructure is not IT/cloud-friendly. Sure, file-based workflows
are relatively mature, but real-time or
live AV systems rely on the SDI streaming
ecosystem for solutions.
In response, industry bodies are developing
best practices and standards for
interoperable methods for transporting in real-time and live; storing, converting and
publishing media assets by leveraging IT
methods. If these groups are successful,
the world of custom AV devices and transport
will dissolve into a world of IT/cloud
services and solutions. The domains of AV-specific
and IT/cloud will co-exist for many
years so, bridging between them will become
a way of life.
DIRECT VS. INDIRECT
The basic challenge is to replace SDI and AES3 audio streaming and associated
ecosystems. SDI/AES3 have reigned
as kings during the past 20 years because
of their excellent lossless, low-latency QoS
and frame-accurate switching.
There are fundamentally two different
approaches for replacements: direct and
indirect. The direct approach streams AV
payloads over Ethernet using LAN Layer 2
(MAC addressing) and/or Layer 3 (IP addressing).
The goal is to replicate the best
features of SDI and its ecosystem (including
point-to-multipoint routing), using Ethernet/
IP. SMPTE standard ST2022-6 maps
SDI payload over IP and was conceived
primarily for point- to-point trunking. It
was not intended as a direct replacement
for SDI but it comes close in some regards.
Frame-accurate video switching using “off-the-
shelf” Ethernet switches still poses the
biggest challenge. Tests and demos have
shown this is possible, but there are no media
industry standards yet.
|Fig. 1: Technology coverage against user requirements from JTF’s request for technology (chart
courtesy of JTF sponsors).
The indirect approach does not attempt
to duplicate SDI over Ethernet. Rather, the
idea is to move sequential chunks of AV
from a sender to receiver using TCP/IP or
other reliable transport protocol. For example,
it’s possible to move one frame, possibly
lightly compressed, (or a few frames or
even K lines) at a time between nodes. This
method uses successive, continuous bulk
transfers and nodal buffering to smooth
out any network issues due to TCP jitter or
latency. Think of this as a real-time quasi-file
transfer between nodes. It’s not isochronous
SDI streaming but it gets the job done.
Sending timed and chunked AV can
maintain real-time transfer equivalence.
The nodal buffering may add a full video
frame of delay per node to keep the video
in frame sync. In a large system, it’s easy to
accumulate 0.5 to 1 second of end-to-end
delay. While not a show stopper for many
applications, this does complicate system
design and “nodal sync” could become the
next “lip sync” issue. One beauty of SDI and
the direct approach is very low node-to-node
delays of less than one line of HD video.
On the other hand, the indirect method
is cloud-friendly since it should operate
fine over lossy, non-deterministic networks.
JOINT TASK FORCE ON
The move to SDI/AES3 replacement
won’t happen overnight but there is
progress in this area. One major activity
is the work of the Joint Task Force (JTF)
on Networked Media. In 2013, this group
was formed by The European Broadcasting
Union, the Society of Motion Picture and
Television Engineers and the Video Services
Forum. This triumvirate provided personnel
and resources to manage the work
effort. A main focus is moving AV streaming
transport to Ethernet, but the effort covers
other themes too.
The JTF published an initial Call for
Participation in April 2013. Next, they collected
136 uncorrelated business-driven
user requirements from many responders
which were boiled down to a correlated
16 “super user requirements” with 122
supporting points. In September the JTF
issued a request for technology (RFT) to
fulfill one or more of the user requirements.
The JTF received 27 submissions
with the respondents submitting a total
of 66 different technologies applied to the
user requirements. Finally, in December,
the JTF published a Gap Analysis paper
which summarizes the 27 submissions
and accompanying technologies, organizing
the information into charts and brief
abstracts. See smpte.org/jt-nm for free access
to the final report.
To no surprise, most of the submissions
covered some, but not all, of the
super user requirements. A few provided
technology references to meet all of the
16 super user requirements. Fig. 1 lists
the 16 main requirements (abbreviated,
see report for details) on the vertical axis
and the average requirements coverage
from all 27 responders on the horizontal
axis. For example, on average, the requirements
for AV streaming (STREAM) were
fully met by about 23 percent of the responders.
This means that some submitted
responses did not cover any of the aspects
for the STREAM user requirements
while others provided much more.
In general, the responses shed light on
several relevant areas and a few technologies
(out of the 66) have clear momentum.
These include; IEEE 1588-V2 (PTP),
SMPTE ST 2022-6, and IEEE 802.1 AVB. The
three were mentioned more often than
other technologies in the submissions.
This effort is a first for our industry. The
report provides a respectable survey of
what vendors and other industry bodies are
thinking along the IT/cloud vein. Stay tuned
to see what the JTF sponsors will do next
now that the report has been published. It
would be a shame to lose the momentum
for this work. Many of the submitted technologies
and similar input will be used to
build the next generation of IT/cloud media
infrastructure. This area is moving fast so
don’t take your eye off the ball.
Al Kovalick is the founder of Media
Systems consulting in Silicon Valley. He
is the author of “Video Systems in an IT
Environment (2nd ed).” He is a frequent
speaker at industry events and a SMPTE
Fellow. For a complete bio and contact information,