Re: [Dots] Questions about draft-doron-dots-telemetry-00

[Flemming Andreasen <[e-mail]>]

On 11/14/16 3:14 AM, Ehud Doron wrote:
>
> Hi Roman
>
> Thanks a lot for your questions.
>
> Before getting in to each of your question, I would like to explain 
> the goal and objectives of our work.
>
> We want to *define and standardize* all the pieces of information 
> needed by DOTS Server mitigation environment for optimal mitigation. 
> This DOTS Telemetry need to be optionally signaled as part of “DOTS 
> Client asks Server for help” various signaling process.
>
> We want to bring the DOTS Telemetry various needs and meanings to be 
> discussed and agreed in the DOTS WG.
>
> After we will all get to the agreement on the various attributes 
> needed, we believe the other DOTS “core” docs need to somehow adopt 
> them and update accordingly.
>
> Essentially we would like to build a "Telemetry" extension standard to 
> the basic DOTS protocol, and this document would define that. Keeping 
> the telemetry specification in a separate draft is good approach for 
> future extensibility.
>
> We expect that the other DOTS draft will benefit this draft by 
> adopting the approaches it suggests to enrich the DOTS signaling. For 
> this goal, we as the DOTS WG first need to better understand (and 
> agree in the group) what we see as baseline mandatory capabilities and 
> what we see as extensions. We are just at the early stages of this 
> work, the current draft defines an initial set of attributes and 
> obviously more discussions in the WG are needed to define the set of 
> attribute that are the “DOTS Telemetry”, and to what extend the other 
> DOTS draft should benefit the DOTS Telemetry.
>
> To your questions:
>
> (1) Is there an undocumented telemetry use case that needs to be added 
> to the use case WG document?  Section 4.0 suggests that.
>
>      Yes. We want to DOTS use cases, the existing and maybe new 
> others, to be enriched with Telemetry related items. It is important 
> to emphasize *“when and where”* DOTS Telemetry can , or should, be 
> signaled such that the mitigation gained for the use case is far 
> better. We want also the other docs (requirements, architecture, 
> protocols, data model and so on) to take the similar approaches.
>
> (2) Is there another protocol draft coming that will incorporate the 
> attributes enumerated in Section 3?  Or are the existing protocol 
> drafts supposed to adopted these?
>
>       As said, the existing protocol drafts supposed to adopt DOT 
> Telemetry approaches by incorporating the attributes enumerated in 
> this draft.
>
> (3) What is the relationship you see between this draft and 
> draft-andreasen-dots-info-data-model-01 (especially given all the 
> authors on the latter are in the former as well)?
>
>       The DOTS Telemetry draft purpose is to *define* the Telemetry 
> attributes needed. The actual data model objects needed to signal 
> these attributes are expected to be define in the info-data model doc. 
> The info-data model draft will define an extension mechanism that 
> supports adding telemetry. To what extent it will also define 
> telemetry is a more open question for the group.
>
To clarify my view on 2) and 3), I think telemetry attributes (with the 
possible exception of a small baseline set) should be defined as 
extensions in separate documents, and that applies to the info-data 
model as well as the protocol. I do think the info-data model as well as 
the protocol should be explicitly defined to accommodate this by 
including an extension model that supports this in a well-defined and 
interoperable manner (modular approach).

Thanks

-- Flemming

> Best wishes,
>
> *Ehud Doron *| Senior Architect, *Radware* CTO office | 
> *M:*[phone] | *T:* [phone]
>
> *From:* Roman D. Danyliw [mailto:[e-mail]]
> *Sent:* Friday, November 11, 2016 6:30 AM
> *To:* Ehud Doron <[e-mail]>; [e-mail]; 
> [e-mail]; Xialiang (Frank) <[e-mail]>; 
> kaname nishizuka <[e-mail]>
> *Cc:* [e-mail]
> *Subject:* Questions about draft-doron-dots-telemetry-00
>
> Hello Ehud, Tiru, Flemming, Frank and Kaname!
>
> Thanks for producing and submitting this draft!
>
> Without getting into the details, what did you have in mind with this 
> draft.  Specifically:
>
> (1) Is there an undocumented telemetry use case that needs to be added 
> to the use case WG document? Section 4.0 suggests that.
>
> (2) Is there another protocol draft coming that will incorporate the 
> attributes enumerated in Section 3? Or are the existing protocol 
> drafts supposed to adopted these?
>
> (3) What is the relationship you see between this draft and 
> draft-andreasen-dots-info-data-model-01 (especially given all the 
> authors on the latter are in the former as well)?
>
> Thanks,
>
> Roman
>