Re: [5gangip] Network tokens draft

The problem that we are solving is that QoS/bearers, zero-rating etc require complex application signatures to map and classify traffic to the desired service (TFT and packet filters in 3GPP-speak). These are very expensive and manual to maintain, they are inaccurate, vulnerable to fraud, bad for privacy, and don't work with upcoming protocols like ESNI. We had 85% failure rate on integrations based on application signatures because of the reasons mentioned above.

The approach we take with network tokens is that traffic differentiation services should be complemented with access management practices, not traffic classification. And network tokens are merely a mechanism to facilitate access management here.

Y.

=====================
Yiannis Yiakoumis
Co-Founder & CEO
https://selfienetworks.com | +1-650-644-7857

On Fri, Jul 10, 2020 at 10:15 AM, Ca By < cb.list6@gmail.com > wrote:

> 
> On Fri, Jul 10, 2020 at 10:03 AM Yiannis Yiakoumis < yiannis@ selfienetworks.
> com ( yiannis@selfienetworks.com ) > wrote:
> 
> 
>> "In mobile, the the bearer is a bit of a virtual circuit.  Each UE has one
>> or more bearers which are assigned an e2e qos policy.  Your cases for
>> tokens would be stronger if you could show how tokens overcome a defined
>> challenge that is not solved today.  It is important for credibility to
>> demonstrate knowledge of the existing approach … otherwise the suggestion
>> appears out of context"
>> We are actually building a prototype that combines tokens with a dedicated
>> bearer for an LTE network, and provides E2E QoS both at uplink and
>> downlink. Implementation happens at https:/ / github. com/ network-tokens/
>> ( https://github.com/network-tokens/ ). Happy to share more details if you
>> are interested.
>> 
>> 
>> 
>> 
>> Y.
>> 
>> 
> 
> 
> 
> But why is this better than the existing bearer / pcrf model?
> 
> 
> What problem are you solving that is novel or better than the existing
> approach?
> 
> 
> Just saying your model is a great is not helpful.
> 
> 
> 
>> 
>> 
>> 
>> =====================
>> Yiannis Yiakoumis
>> Co-Founder & CEO
>> https:/ / selfienetworks. com ( https://selfienetworks.com ) |
>> +1-650-644-7857
>> 
>> 
>> 
>> On Fri, Jul 10, 2020 at 9:35 AM, Yiannis Yiakoumis < yiannis@ selfienetworks.
>> com ( yiannis@selfienetworks.com ) > wrote:
>> 
>>> subscribed late and can't reply inline to follow-up comments. Short
>>> response to Ca By's point on whether the network trusts the UE.
>>> 
>>> 
>>> 
>>> "A fundamental principle of mobile qos is that you do not trust the UE, as
>>> it can be compromised to make all traffic high priority. Only the network
>>> can reliably and securely allocate resources based on policy."
>>> The basic architecture for network tokens is agnostic on the trust
>>> relationships between UE, app provider, network and server. It provides
>>> mechanisms to encrypt/sign a token, and metadata to prevent replay and
>>> spoofing attacks so that operators can adjust it to the appropriate trust
>>> model. It borrows a lot from the ideas implemented in JWT, CWT and OAUTH2.
>>> 
>>> 
>>> 
>>> 
>>> Yiannis
>>> 
>>> 
>>> 
>>> 
>>> 
>>> =====================
>>> Yiannis Yiakoumis
>>> Co-Founder & CEO
>>> https:/ / selfienetworks. com ( https://selfienetworks.com ) |
>>> +1-650-644-7857
>>> 
>>> 
>>> 
>>> On Thu, Jul 09, 2020 at 10:00 AM, Tom Herbert < tom@ herbertland. com (
>>> tom@herbertland.com ) > 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 (
>>>> 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 (
>>>> 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
>> 5gangip@ ietf. org ( 5gangip@ietf.org )
>> https:/ / www. ietf. org/ mailman/ listinfo/ 5gangip (
>> https://www.ietf.org/mailman/listinfo/5gangip )
> 
> 
>