White Papers

Introduction
Substantial changes will take place in information and communications
technology as we migrate to the next generation compute platform.
This platform will likely include elements of several adjacent technologies.
These adjacent technologies include 5G Networks, Edge Cloud Computing,
IoT, AI and Machine Learning, and Augmented and Virtual Reality (AR/VR).
As these future technologies mature, new and additional workload
requirements will be placed on the information and technology platform to
support these new areas of focus.
In an evolving edge compute architecture paradigm, increasing concurrency of
workloads creates constrained environments across existing standard
connectivity interfaces, while the requirement to increase performance and
reduce latency will increase. These needs are further exacerbated by the need
for backward compatibility with existing interfaces and conformance to current
standards.
Forward and Back Channel
Traditional content consumption models, whether broadcast/multicast or
streaming, are largely unidirectional: the primary transmission originates at a
host, or sources, and terminates at an end-point sink, or sinks. The direction of
this primary transmission from source to sink is typically referred to as the
forward channel. An example of forward channel over a network would be
streaming of YouTube content from a YouTube https server (source) to an
edge client browser or app (sink). Example of a local forward channel are the
video, audio, and control output data from an IHS tethered wired or wirelessly
with an XR (AR or VR) HMD.
In addition to the forward channel, real-time interactive content experiences
generally require the sink to respond with control, data or other ancillary
information. The direction of transmission from a sink back to a source is
typically referred to as a back channel.
Forward channel transmissions, by themselves, face constraints. In general,
the desire to maximize the content transmitted leads to pushing the bandwidth
limitation of the communications channel. Numerous techniques are used,
including compression, resiliency and error concealment, to maximize amount
of content transmitted within transmission bandwidth limitations of the forward
channel, while eliminating or minimizing loss or error in environments where
transmission bandwidth is constrained or faulty.
For traditional transmissions, the backchannel requirements are usually less
than forward channel, in terms of overall bandwidth. However, emerging user
experiences, such as Augmented or Virtual Reality, place much greater
demand on the backchannel, both in terms of amount of data transmitted and
synchronized, to a greater need to reduce latency, or increase responsiveness,
of the transmission. The forward and back channel communication, in fact,
combine to define an overall requirement for round-trip latency, for many use
cases.
A straightforward example of this requirement is an AR or VR head mounted
device (HMD) for use with a host or end user system. The forward channel
transmission is comprised of audio, video and control information for presenting