Re: [5gangip] Network tokens draft

[Alexandre Petrescu <[e-mail]>]

Le 10/07/2020 à 16:44, Tom Herbert a écrit :
> On Fri, Jul 10, 2020 at 7:40 AM Behcet Sarikaya
> <[e-mail]> wrote:
>> 
>> Tom,
>> 
>> It has already been announced that this list is no longer dealing
>> with 5G issues. We decided to move to 6g, please check recent
>> mails.
>> 
>> As such your topic is no longer of interest and we invite you to
>> join us into the new and exciting area of 6G.
>> 
> 
> Very well, if people are interested in further discussion please
> post on the Network-tokens list.

Or we can think that any new proposed technology could fall under the
ummbrella of 6G rather than that of 5G.  Because 5G is no longer under
development, I think.

By the way, if I wanted to implement a network token concept I would
consider the use of existing Flow Labels field in the IPv6 header.  A
flow label would contain a mapped value of a token.

If I wanted a Network Token to be for an endpoint on a 6G link, then I
would look at what is the 6G link access method of the PHY of the UE.
MAybe that link access method (an *-MA method, a Multiple Access method,
like in CDMA, OFDMA, SCDMA, etc.) could use a Network Token.

And then, maybe putting IP on that 6G link would need a mapping between
a Flow Label and a network token in the *-MA method.

Alex

> 
> Tom
> 
>> Behcet
>> 
>> On Fri, Jul 10, 2020 at 9:28 AM Tom Herbert <[e-mail]>
>> wrote:
>>> 
>>> On Fri, Jul 10, 2020 at 5:10 AM John Grant <[e-mail]>
>>> wrote:
>>>> 
>>>> This adds even more per-packet processing and makes headers
>>>> even bigger. In ETSI ISG NIN
>>>> https://www.etsi.org/technologies/non-ip-networking we're
>>>> working on reducing them, as mobile operators (and no doubt 
>>>> others) want, by getting all this information into the control
>>>> plane where it belongs.
>>>> 
>>> John,
>>> 
>>> Network tokens provide the API between an application on a UE and
>>> the network for the application to specify exactly the service it
>>> wants. At the first hop in the network, the device can compress
>>> the headers, use MPLS, put into some non-IP format, whatever it
>>> wants as long as the user gets the service they've requested and
>>> correct protocol behavior is maintained. If the operator wants to
>>> create a state for every single flow in their network that's
>>> their prerogative, but good luck scaling that to hundreds of
>>> billions of flows and getting it to work seamlessly across mobile
>>> events-- the "connectionless" model of IP replaced
>>> "circuit-switched" networks for good reason!
>>> 
>>> Tom
>>> 
>>>> -- John Grant Nine Tiles, Cambridge, England [phone]
>>>> and [phone] http://www.ninetiles.com
>>>> 
>>>> 
>>>> On 09/07/2020 18:00, Tom Herbert wrote:
>>>>> This is a draft on "Network Tokens" which is of relevance to 
>>>>> facilitate fine grained QoS in 5G networks.
>>>>> 
>>>>> https://tools.ietf.org/html/draft-yiakoumis-network-tokens-01
>>>>>
>>>>>
>>>>> 
There is also a mailing list in
>>>>> https://www.ietf.org/mailman/listinfo/network-tokens
>>>>> 
>>>>> We are planning to present in tsvwg and app aware networking
>>>>> and possibly have a side meeting on this topic in IETF108.
>>>>> 
>>>>> Thanks, Tom
>>>>> 
>>>>> _______________________________________________ 5gangip
>>>>> mailing list [e-mail] 
>>>>> https://www.ietf.org/mailman/listinfo/5gangip
>>>> 
>>>> _______________________________________________ 5gangip mailing
>>>> list [e-mail] 
>>>> https://www.ietf.org/mailman/listinfo/5gangip
>>> 
>>> _______________________________________________ 5gangip mailing
>>> list [e-mail] 
>>> https://www.ietf.org/mailman/listinfo/5gangip
> 
> _______________________________________________ 5gangip mailing list 
> [e-mail] https://www.ietf.org/mailman/listinfo/5gangip
>