[Nmc-wg] 2/17 Call Canceled

[Jeff W. Boote]

Hi Roman,

On Feb 17, 2011, at 4:32 AM, Roman Łapacz wrote:

> W dniu 2011-02-17 06:39, Jeff W. Boote pisze:
>> Roman,
> 
> Hi Jeff,
> 
>> We intend to have a version of the circuit-monitoring softare available for our first DYNES deployments in April. We are having to make choices in our implementation work for some of the things we are currently discussing in this group now. It is our intention to move to whatever is agreed to over time.
>> 
>> Basically, I"m not all that concerned with most of the syntax issues that have come up. I think those can be handled by software/protocol versioning. However, I am concerned with access and interaction pattern discussions. Those things are more difficult to modify. This is one of the reasons I was attempting to keep our discussion very focused on the protocol interaction. Also... the 'shortcut' discussion that has happened seemed a little strange to me. I think we should design the protocol for the most general case first. If there are ways it can be optimized for certain cases, I believe that should happen later.
> 
> That 'shortcut' discussion you mean the discussion on the MAs with fully aggregated data and topology (Inter-Domain View in the spec doc v0.20)? Mind that this approach has been proposed for the AutoBAHN and pS integration and as an input for our discussion on more general solution. As you can see I'm not stuck to things originally proposed by GN team. For example, the schema for descriptors have changed a lot and that's fine because we've had a fruitful discussion and a good progress (Aaron, Freek, Jason, correct me if I'm wrong).

Yes, I agree the discussion has been very fruitful. I believe much of the differences that we are coming up with in the access patterns have to do with the fact that AutoBAHN is more of a single service solution, while we have been thinking of things in a more distributed fashion. For example, not all of our topology data is necessarily located in a single service. (As has been pointed out in the other thread.)

> Jeff, your comment about focusing on access and interaction pattern was important because we forgot a bit about the client side while working on descriptors. But I think the recent discussion on this is going in a good direction. If you think it's not please send more details. Your input has been always precious and helped a lot many times.
> 
> About the timetable, can we agree that
> - by the end of Feb we should have a stable version of schema for descriptors (examples and RNC file with definitions; the discussion on NML is important here so Freek please comment that if you see some problems),
> - by the mid of next month we should have the client access to topology and measurement data defined (sequence diagram(s) and xml message examples)?
> Any other proposal?

This timetable sounds fine. I appreciate very much that you are escalating things for our benefit.

jeff


> 
> Thanks,
> Roman
> 
> 
>> thanks,
>> jeff
>> 
>> On Feb 15, 2011, at 8:40 AM, Roman Łapacz wrote:
>> 
>>> W dniu 2011-02-15 16:17, Jason Zurawski pisze:
>>>> Hi All;
>>>> 
>>>> I will be unavailable for our regular call this Thursday so I propose we
>>>> cancel.  If anyone would like to go on (to talk about pending work on
>>>> Circuit Monitoring for example), the MCU should still be available.
>>> Hi,
>>> 
>>> I wanted to talk about dates. This is important because few days ago I
>>> had a discussion with Ann Harding and she asked me to talk with you
>>> about the timetable of our work and its progress. She insisted to have
>>> the milestones to track whether we progress in the right pace and have
>>> the results.
>>> 
>>> Aaron, I know that you should have the implementation quite soon. Can
>>> you tell me when? Can we choose a dates when we should have the schema
>>> for descriptors and architecture (or client access/client lookup
>>> procedure) completed (at least the versions that are quite mature). I
>>> think that the format of descriptors is quite mature (although I haven't
>>> read your last discussion with Freek; tomorrow will do that). It seems
>>> that the client access requires still some discussion. What's your
>>> proposal for that timetable? Sync that with your implementation work.
>>> 
>>> Ann told me that the running GN3 implantation (integration of AutoBAHN
>>> and pS) should be ready by Sept.
>>> 
>>> Cheers,
>>> Roman
>>> 
>>>> Thanks;
>>>> 
>>>> -jason
>>>> _______________________________________________
>>>> Nmc-wg mailing list
>>>> Nmc-wg at ogf.org
>>>> http://www.ogf.org/mailman/listinfo/nmc-wg
>>> _______________________________________________
>>> Nmc-wg mailing list
>>> Nmc-wg at ogf.org
>>> http://www.ogf.org/mailman/listinfo/nmc-wg
>