Re: [EToSat] about LOOPS(Localized Optimization of Path Segments) and etosat

Dear All,

On Thu, Jan 3, 2019 at 1:46 PM Marie-Jose Montpetit <marie@mjmontpetit.com>
wrote:

> The files also refer to the nwcrg on-going work. It is a common practice
> to run coded (FEC) sessions over tunnels to isolate the effects of coding
> across paths. And we have work on nw coding over satellite also. Would this
> group be open for me to send this thread over to our list?
>
> Also in the COIN proposed RG we are also looking at per hop optimization
> but in the context of in-network computing (and that includes FEC,
> transcoding, pre-fetching, data decompostion etc.) which I can see are also
> related. So I could copy there too.
>
> Thanks for copying me.
>
> mjm
>
>
> Marie-Jose Montpetit, Ph.D.
> mariejo@mit.edu
> marie@mjmontpetit.com
> +1-781-526-2661
> @SocialTVMIT
>
>
>
> On Jan 3, 2019, at 12:48 PM, Border, John <John.Border@hughes.com> wrote:
>
>
>
> Clearly LOOPS should not focus only on satellite.  But, I think LOOPS is
> directly relevant to ETOSAT.  Techniques like LOOPS are likely to be a key
> part of the solution to the performance over satellite problem.
>
> I will be sending out an email to ETOSAT in the next couple of days to try
> to kickstart discussion overall.
>
>
> John
>
>
>
> -----Original Message-----
> From: Carsten Bormann <cabo@tzi.org>
> Sent: Friday, December 07, 2018 5:59 PM
> To: Border, John <John.Border@hughes.com>
> Cc: Liyizhou <liyizhou@huawei.com>
> Subject: Re: about LOOPS(Localized Optimization of Path Segments) and
> etosat
>
> WARNING: The sender of this email could not be validated and may not match
> the person in the "From" field.
>
> CAUTION: This email originated from outside of the organization. Do not
> click links or open attachments unless you recognize the sender and know
> the content is safe.
>
>
> WARNING: The sender of this email could not be validated and may not match
> the person in the "From" field.
>
> Hi John,
>
> On Dec 7, 2018, at 10:37, Liyizhou <liyizhou@huawei.com> wrote:
>
> I copied to Carsten who was leading the side meeting. He might provide
> more information.
>
>
> As Yizhou said, there are no formal minutes of this side meeting, but I
> did record my impressions in the attached short summary.  I have also
> attached the slides that we used, as the summary may not be comprehensible
> without those.
>
> Given that Spencer specifically asked us to discuss this in the context of
> etosat, it would certainly help to know what the intentions were in
> creating this list and to speculate a bit how a potential etosat activity
> might cross-pollinate with LOOPS.
>
> Grüße, Carsten
> <Summary of the IETF103 LOOPS side meeting, 2018-11-06.pdf><LOOPS.pdf>
>
>
> _______________________________________________
> EToSat mailing list
> EToSat@ietf.org
> https://www.ietf.org/mailman/listinfo/etosat

My intention for creating this mailing list was as a first step toward
working through the performance implications of encryption in situations
where end-to-end performance mechanisms may encounter difficulty because of
path characteristics, which (thinking out loud) could involve one
problematic hop, or multiple problematic hops, and non-end-to-end
mechanisms aren't able to assist now, if traffic is encrypted.

(Note - this was in the context of comparing HTTP over QUIC to HTTP over
TCP, and we all realize (I hope that TCP stacks have decades of hand
optimization and QUIC implementations do not. My intuition is that even
with tuning, QUIC may be at a disadvantage on paths containing satellite
links, so there is a real underlying problem to be solved)

I created this list at John's request during IETF 101, and John's concern
was specifically related to paths containing at least one satellite link,
because we know that's broken.

At IETF 103, I started thinking that LOOPS was facing the same problem, in
an entirely different application, so believe it is appropriate to use this
mailing list to explore what specific technical challenges the two sets of
proponents have in common. And I suspect the satellite community and the
LOOPS community may not be the only two communities with similar problems
in this space.

I have guesses about where this discussion might go, but for now, I'll let
people continue to work together. Make good choices!

Spencer