Re: [6gip] Minutes taken by Dirk

[Alexandre Petrescu <[e-mail]>]

Le 28/07/2020 à 19:20, Behcet Sarikaya a écrit :
> Hi all,
> 
> On Tue, Jul 28, 2020 at 8:46 AM Behcet Sarikaya
> <[e-mail] <mailto:[e-mail]>> wrote:
> 
> Here are Dirk's notes from our side meeting:
> 
> Minutes of side meeting 6Gip during IETF#108 online – July 27, 2020 
> 16h UTC About 20 people had dialed in. After some technical issues
> (not with Meetecho but Webex) Behcet went over the slides posted on
> the ML. Alex asked in the chat for clarification what Mandate (in 
> Mandate-Driven Architecture) does mean. Behcet replied that ”mandate”
> is seen as collection of network services required from underlying 
> network(s) to comply with dynamic end-to-end QoS needs from specific 
> applications (in the WP) Lars complained in the chat: so if the
> proponents of this effort are planning on asking people to make time
> on their schedules for another side meeting at some time in the
> future, *please* have some actual content planned for presentation
> and discussion. Behcet agreed: OK Lars Colins key feedback was that
> IRTF does not work on overarching architecture but particular
> subtopics and problem statements – AI is already worked on in NMRG
> and computing in COIN and a contribution/collaboration there
> regarding 6G should be investigated. Alex tried to answer on what 6G
> is – perhaps (at least) true IPv6 but much more: futuristic use
> cases and scenarios, eventually what will be mandatory use cases? 
> Lars said that it was not clear before the meeting what we were 
> asking for. Alex mentioned that successful RGs have started low – but
> Lars emphasized that they dealt with a well scoped problem. Colin
> said that development from 5G to 6G is no concrete focus, not worked
> on by existing RGs. Behcet argued that network management is also no
> real focus. Dirk tried to explain how MDA could serve here. Marco
> asked which term could be moving towards a solution. It should be
> worked on gaps between SBA++ and current enhanced SBA (Rel. 16/17). 
> Behcet mentioned future use cases and Alex clarified that hologram 
> use case mentioned in WP requires much bandwidth.
> 
> 
> 
> Alex had mentioned a few requirements/ use cases like self driving
> cars which is not captured in the minutes.
> 
> Alex can you please remind them?

When people talk about which applications are really in a need of a new
network access like 5G, several answers are given, but typically the
self-driving cars come always in the list.

Recently, the talk of use-cases of 5G seems to be more and more focused
on this 5G for self-driving cars much more than 5G for anything else.
Yet they do ignore significant doubts.

Some sub-use cases under the generic umbrella of the topic of
self-driving cars are:

- tele-operation: the ability of a human operator situated in a remote
control center to take control and guide a self-driving car, by using a
very reliable (ultra) 5G link.  This has itself several sub-sub-cases
like: assisting the manoeuvres of cars, bringing a car from parking to
user, parking a car, moving the car away from an arriving emergency
convoy, stop the movement of a car-gone-crazy, and more.

- 5G-V2X: the applications designed currently for other link medias like
802.11-OCB at 5.9GHZ (DSRC,ITS-G5) are here performed on a 5G link media
instead.  One sub-sub-case is, for example, advertising the presence of
a car by broadcasting a 'CAM' message (Cooperative Awareness Message).

- others, like low-latency communication between self-driving cars and
traffic light controller units, via a Control Center; this is to display
the color of the traffic lights on the car's dashboard even when the
car's video camera does not see (fog, etc.)

For each of these 5G sub-usecases for self driving cars there are also
serious doubts, that I could mention separately.

Looking onward to 6G, one should wonder why these use-cases for 5G and
self-driving cars do not materialize on a wide scale.  Answering that
question might give a hint to what 6G needs to be.

Alex

> 
> Behcet
> 
> Lars: use case done meant problem statement (?) and reminds that 
> IntServ had not been successful and it would be good to understand
> whether to re-invent old issues. Toerless thiks that similar
> technology could at other times be successful. L2 TSN (IEEE) using
> detnet concept (used in limited campus scenarios) could be a good
> exercise ? use cases and business models are the difficult parts (for
> all new technologies/approaches) Colin asked how to fit to new
> proposals and Behcet answered to use this opportunity to discuss 
> more. Hannu emphasized that 6G would be a new generation compared to
> 5G – THz technology as good example. We should look at use cases 5G
> is not able to do – very tricky thing! Alex mentioned a typical use
> case which is not deployed in 5G: V2X (neither network-based nor 
> direct communication) - more 5G use cases (other than eMBB) may not 
> be deployed in 5G but only with 6G (as with other specifications in
> 3GPP never deployed).
> 
> Hannu expects that in 6G the same services may be deployed, but in a 
> simpler manner. Dirk agreed – also more sustainable methods could
> help deployment. FMC is another example already standardized for 3G
> but coming actually with 5G now. Alex asked whether people active in
> 3GPP could confirm that 6G has already been discussed before? Tim
> said that 3GP does not yt speak of 6G specifically but NGMN (Next
> Generation Mobile Networks association) has started to talk on this
> (WP#2). The chairs thanked all participants for joining and
> contributing to good discussion to be continued on the list and on
> future side meetings.
> 
>