I am the assigned ArtArt reviewer for this draft.
Document: draft-ietf-mops-ar-use-case-15
Reviewer: Paul Kyzivat
Review Date: 2024-04-01
IETF LC End Date: 2024-04-02
IESG Telechat date: ?
Summary:
This draft is on the right track but has open issues, described in the
review.
General Comments:
This review is from a reviewer with no prior knowledge of this field. It
is based solely on the text of the document and a few of the references
I was able to access. ([EDGE_*] and [XR]) This was challenging because
there are a lot of references and many of them are not readily accessed
online. I generally found that some key terms are not defined in the
document. I thought perhaps they were defined in the references, but in
the cases I was able to check that wasn't the case.
ISSUES: 3
NITS: 1
Issues and Nits:
1) ISSUE: Audience and purpose
The abstract states: "The intended audience for this document are
network operators who are interested in providing edge computing
resources to operationalize the requirements of such applications."
Is this true? This doc is a product of the MOPS WG. The charter of MOPS
is broader than that. The MOPS charter says:
"MOPS provides a venue for both video industry and Internet engineering
experts to engage in discussion of video technology’s requirements of
networking standards, as well as proposals for new uses of IP technology
in video. Where new protocols are needed, MOPS will help identify
candidate venues for their development."
It seems to me that this implies a broader target audience. This doc
seems to be intended to expose the membership of MOPS to the content of
[XR].
I suggest you reconsider and carefully define the intended audience and
purpose of the document. And then ensure the document is consistent with
that.
2) ISSUE: Where is the Edge?
The Intro says: "Consequently, in order to run applications with XR
characteristics on mobile devices, computationally intensive tasks need
to be offloaded to resources provided by Edge Computing."
This is a definitive statement but lacks substantiation. You do make the
case that some tasks need to be offloaded to *somewhere* "closer" than
the cloud. But because there is no definition of "Edge Computing" it is
impossible to judge if that is an appropriate place.
The following seems to further define what you mean:
"This document discusses the issues involved when edge computing
resources are offered by network operators to operationalize the
requirements of XR applications running on devices with various form
factors." But "network operator" is not defined. Is it focused on
commercial regional providers of telecom/network services to businesses
and individuals, or does it also cover the operators of LANs deployed in
businesses and individual homes, or what?
The rest of the paper makes assumptions about the characteristics of
this environment. But those characteristics are never stated. This
document offers no conclusions. Since it is focused on resources
provided by network operators I would like to see some conclusions about:
- how close provider edge computing resources could be and what latency,
bandwidth, and jitter characteristics could be possible.
- which if any of the use cases are feasible using network operator
resources.
Alternatively the document could drop the focus on network operator
resources and instead focus on how the needs of the use cases constrain
where the resources can be deployed.
3) ISSUE: References
It's good that there are many references. Several of the references that
I tried were not readily available online for free. So I didn't
investigate them. This was an impediment to understanding the document.
Are the readers expected to be familiar with all this material?
Please check that the document isn't relying on reader knowledge of
references that may be inaccessible to the target audience.
4) NIT: Run-on sentence
These is an extreme run-on sentence in the Intro:
"In particular, the edge devices deploy ... disaggregation ...
virtualization ... commoditization ..."
This needs a thoughtful rewrite and breakup into multiple sentences and
possibly bullet points.
--
last-call mailing list
last-call@xxxxxxxx
https://www.ietf.org/mailman/listinfo/last-call
