Re: [alto] Cellular information exposure discussion (Fwd: New Version Notification for draft-huang-alto-mowie-for-network-aware-app-04.txt

[Zili Meng <[e-mail]>]

Hi Richard,

Thank you for the feedback in detail! We will start to investigate the
related work (e.g., MOWIE, use cases, Yaxiong's W1 paper, etc.) and try to
write such a survey soon.

On Tue, Jul 19, 2022 at 3:52 PM Y. Richard Yang <[e-mail]>
wrote:

> Hi Zili,
>
> Thanks for the quick feedback! Please see below.
>
> On Tue, Jul 19, 2022 at 12:18 PM Zili Meng <[e-mail]> wrote:
>
>> Hi Richard and all,
>>
>> Thanks for bringing this up!
>>
>> Sure we'll be very happy if we could contribute to the IETF and the alto
>> group on how to feedback the network conditions (e.g., cellular
>> information) to the application (the sender). And I also feel that a survey
>> and a case study would be very helpful to sort out what we need in the
>> scope of alto.
>>
>
> It looks that we are in agreement :-) I did a preliminary search, but do
> not see a systematic survey. Such a survey can be an ietf work or an
> independent work as a paper. We can wait for the guidance from the WG
> chairs or AD.
>
> As a starting point, I can see that the survey should include not only the
> distribution methods (D-* in the previous email), which are potential
> solutions, but also
> - a list of benchmarking settings, including triggering events, such as
> the bandwidth fluctuations that your work addresses;
>
> - a list of use-case restriction or required features, such as
> authentication (restriction), and carrier aggregation, which motivates the
> W1 work;
>
> - a list of performance metrics and their requirements, including those
> application metrics and requirements discussed in MOWIE, and network
> metrics (e.g., active users) and the requirements on their delivery by the
> information channel.
>
>
>> Unfortunately I may not be able to attend the alto meeting in IETF 114.
>>
>
> No problem at all. Email is a good starting point.
>
> Do you feel it would be better if we first prepare some writeups or drafts
>> for the C2 (send the network conditions back to the sender)? We could then
>> bring writeups back on the design space and / or the use cases so that we
>> can have something to discuss with.
>>
>
> Yes. See above. Let’s write the draft soon.
>
> Richard
>
>>
>
>
>> A preprint of the W2 paper is available at
>> https://zilimeng.com/papers/zhuge-sigcomm22.pdf for your information.
>> We'll also be highly appreciated if there are any comments or feedback.
>>
>> On Tue, Jul 19, 2022 at 1:46 PM Y. Richard Yang <[e-mail]>
>> wrote:
>>
>>> Hi ALTO WG,
>>>
>>> During the weekly meeting today, the design team discussed about the
>>> document, and suggested that the authors update the WG on a new version of
>>> the MOWIE draft, which was uploaded last week and focuses on cellular
>>> network information exposure to network-aware applications. The links to
>>> the new version and the diff showing the updates can be found below. Any
>>> comments, feedback, or interests in working together are welcome and
>>> greatly appreciated.
>>>
>>> Using this email, we also want to include that there are two quite
>>> relevant pieces of work:
>>> W1. PBE-CC: Congestion Control via Endpoint-Centric, Physical-Layer
>>> Bandwidth Measurements (https://arxiv.org/abs/2002.03475), by Yaxiong
>>> Xie, Prof. Jamieson, and Prof. Rexford;
>>> W2. Achieving Consistent Low Latency for Wireless Real-Time
>>> Communications with the Shortest Control Loop (paper to be made public
>>> soon), by Zili, and Prof. Mingwei and team.
>>>
>>> It helps to use this email to point out a bigger picture of the problem
>>> space and related work. One perspective is that the problem space consists
>>> of 3 components:
>>> C1. What cellular-network information should be collected to be sent to
>>> the applications?
>>> C2. How is the cellular information sent to the applications?
>>> C3. How do the applications use the information?
>>>
>>> It is important to note that the full exploration of the problem space
>>> is not included in the current charter. However, this is an important
>>> problem space and hence it helps  to keep track of the progress and
>>> potential impacts on ALTO.
>>>
>>> For C1, the impact will be the list of performance metrics. For C3, the
>>> current ALTO charter does not include items to define application
>>> behaviors, but it can be a point of discussion related to deployment.
>>>
>>> The most relevant component is C2. We discussed the following design
>>> points for C2:
>>> D-broadcast: network broadcasts its state
>>> D-pass-bounce: network marks its state on pass-through packets and the
>>> receiver bounces the state back to the sender
>>> D-direct-send: network sends its state to app directly
>>> Here by state it can be transformed state.
>>>
>>> Current ALTO is designed as D-direct-send. Due to lacking of deployment
>>> of D-direct-send in cellular network, aforementioned W1 uses D-broadcast;
>>> the D-pass-bounce design need at least a round trip time and couples
>>> network feedback flow with data flow, leading to the aforementioned W2 work.
>>>
>>> Many of us feel that the time is ready for introducing a highly
>>> efficient, flexible channel for network state exposure to applications, led
>>> by IETF, and the ALTO team can be a good starting point. As a first step, a
>>> systematic evaluation, which (1) surveys the design space and (2)
>>> introduces the use cases/benchmarking scenarios, can be a good starting
>>> point.
>>>
>>> If there is time available in the coming 114 meeting, we can discuss
>>> more during the presentation. Otherwise, on-list discussion is a good
>>> starting point. We will follow up with more analysis on the list soon but
>>> use this email to get the conversation started.
>>>
>>> Thanks,
>>>
>>> Richard on behalf of Tuesday meeting team
>>>
>>>
>>>
>>> ---------- Forwarded message ---------
>>> From: <[e-mail]>
>>> Date: Mon, Jul 11, 2022 at 2:14 PM
>>> Subject: New Version Notification for
>>> draft-huang-alto-mowie-for-network-aware-app-04.txt
>>> To: Y. Richard Yang <[e-mail]>, Gang Li <
>>> [e-mail]>, Sabine Randriamasy <
>>> [e-mail]>, Yixue Lei <[e-mail]>,
>>> Yuhang Jia <[e-mail]>, Yunbo Han <[e-mail]>,
>>> Yunfei Zhang <[e-mail]>
>>>
>>>
>>>
>>> A new version of I-D, draft-huang-alto-mowie-for-network-aware-app-04.txt
>>> has been successfully submitted by Y. Richard Yang and posted to the
>>> IETF repository.
>>>
>>> Name:           draft-huang-alto-mowie-for-network-aware-app
>>> Revision:       04
>>> Title:          MoWIE for Network Aware Applications
>>> Document date:  2022-07-11
>>> Group:          Individual Submission
>>> Pages:          25
>>> URL:
>>> https://www.ietf.org/archive/id/draft-huang-alto-mowie-for-network-aware-app-04.txt
>>> Status:
>>> https://datatracker.ietf.org/doc/draft-huang-alto-mowie-for-network-aware-app/
>>> Htmlized:
>>> https://datatracker.ietf.org/doc/html/draft-huang-alto-mowie-for-network-aware-app
>>> Diff:
>>> https://www.ietf.org/rfcdiff?url2=draft-huang-alto-mowie-for-network-aware-app-04
>>>
>>> Abstract:
>>>    With the quick deployment of 5G networks in the world, cloud-based
>>>    interactive applications (services) such as cloud gaming have gained
>>>    substantial attention and are regarded as potential killer
>>>    applications.  To ensure users' quality of experience (QoE), a cloud
>>>    interactive service may require not only high bandwidth (e.g., high-
>>>    resolution media transmission) but also low delay (e.g., low latency
>>>    and low lagging).  However, the bandwidth and delay experienced by a
>>>    mobile and wireless user can be dynamic, as a function of many
>>>    factors, and unhandled changes can substantially compromise the
>>>    user's QoE.  In this document, we investigate network-aware
>>>    applications (NAA), which realize cloud based interactive services
>>>    with improved QoE, by efficient utilization of a solution named
>>>    Mobile and Wireless Information Exposure (MoWIE).  In particular,
>>>    this document demonstrates, through realistic evaluations, that
>>>    mobile network information such as MCS (Modulation and Coding Scheme)
>>>    can effectively expose the dynamicity of the underlying network and
>>>    can be made available to applications through MoWIE; using such
>>>    information, the applications can then adapt key control knobs such
>>>    as media codec scheme, encapsulation, and application layer
>>>    processing to minimize QoE deduction.  Based on the evaluations, we
>>>    discuss how the MoWIE features can define extensions of the ALTO
>>>    protocol, to expose more lower-layer and finer grain network
>>>    dynamics.
>>>
>>>
>>>
>>>
>>> The IETF Secretariat
>>>
>>>
>>> --
>>> Richard
>>>
>>
>>
>> --
>> Cheers,
>> ----------------
>> Zili Meng
>> Web: https://zilimeng.com
>>
> --
> Richard
>


-- 
Cheers,
----------------
Zili Meng
Web: https://zilimeng.com