Re: [EToSat] New Version Notification for draft-lhan-problems-requirements-satellite-net-02.txt

IP relies on address aggregation and topology abstraction to handle massive
scales. A link state protocol such as OSPF or ISIS also employs hierarchy
to handle scalability. An OSPF area for example rarely goes beyond 1000
nodes. There are several reasons for that, one of which of course is to
avoid broadcast storms arising from the link events.

A link state protocol with a single area covering an entire small-size
constellation (essentially a flat network with no hierarchy) would perhaps
be able to handle the link events (within the satellite network) and move
traffic between IP gateways on the ground (not necessarily equal to ground
stations) with a reasonable convergence time. However, I don't see how an
OSPF-like hierarchy can ever be configured and maintained in large
satellite networks. The movement alone will turn this into a daunting task.

Those of us with experience in building networking, routing, and forwarding
systems would never expose a large sum of routes to the satellite network
or any other inner network for that matter.
I am not sure what gave the impression that the complexity should be
placed anywhere else BUT the ground.

My point is that link state routing (doesn't mean IP) might work in small
constellations, but I don't think it will be able to handle routing and
forwarding in massive constellations. We should investigate alternatives to
existing terrestrial routing protocols to handle routing and forwarding in
satellite networks.

Trust me, I have been in networking/routing research over the past 30 years
and have built numerous systems that are actually deployed around the world
:)

On Wed, Feb 23, 2022 at 9:57 AM Marie-Jose Montpetit <marie@mjmontpetit.com>
wrote:

> Iridium?
>
> This is from the 90s…
>
> "IP good enough” - what do you mean? There are millions (billions) of
> nodes on the internet.
>
> Knowing the complexity of a satellite segment believe me you want to keep
> the complexity on the ground :)
>
> I think you guys need to do some research :)
>
> mjm
>
> Marie-José Montpetit, Ph.D.
> marie@mjmontpetit.com
>
>
>
> From: Arashmid Akhavain <arashmid.akhavain@gmail.com>
> <arashmid.akhavain@gmail.com>
> Reply: Arashmid Akhavain <arashmid.akhavain@gmail.com>
> <arashmid.akhavain@gmail.com>
> Date: February 23, 2022 at 9:54:27 AM
> To: David Fernández <davidfdzp@gmail.com> <davidfdzp@gmail.com>
> Cc: etosat@ietf.org <etosat@ietf.org> <etosat@ietf.org>
> Subject:  Re: [EToSat] New Version Notification for
> draft-lhan-problems-requirements-satellite-net-02.txt
>
> If I am not mistaken, Iridium and OneWeb constellations are less than 1000
> satellites in size. IP might be good enough for those constellations.
>
> Best regards,
> Arashmid
>
> On Wed, Feb 23, 2022 at 5:34 AM David Fernández <davidfdzp@gmail.com>
> wrote:
>
>> Dear all,
>>
>> In Starlink, a traceroute command does not show the satellites, just
>> the gateway. I think that this is an indication that they are not
>> using IP routing protocols.
>>
>>
>> https://www.reddit.com/r/Starlink/comments/hbicej/maybe_dumb_question_what_will_we_see_on_traceroute
>>
>> https://www.reddit.com/r/Starlink/comments/jo4fcu/traceroutemtr_anyone
>>
>> https://www.reddit.com/r/Starlink/comments/mym6tt/puzzeling_trace_route
>>
>> Has anybody tried traceroute on Iridium? OneWeb? Others?
>>
>> Best regards,
>>
>> David
>>
>> _______________________________________________
>> EToSat mailing list
>> EToSat@ietf.org
>> https://www.ietf.org/mailman/listinfo/etosat
>>
> _______________________________________________
> EToSat mailing list
> EToSat@ietf.org
> https://www.ietf.org/mailman/listinfo/etosat
>
>