[Nsi-wg] Request for developer for NSI visualisation software
Takatoshi Ikeda
ikeda at ote.kddi.com
Wed Oct 5 21:29:37 CDT 2011
Kudoh-san, Jerry-san and all.
It's also good to me.
I'm not familiar with NSI query() so I thought it might take much
time to implement it to Automated Earth.
However, the proposed architecture is easier for me.
If the app on GoogleAppEngine provides all status with HTTP, SOAP
or such a way, I can focus on the development of visualization part.
Thanks,
-Takatoshi Ikeda
On Wed, 05 Oct 2011 14:18:41 -0400
Jerry Sobieski <jerry at nordu.net> wrote:
> This sounds good to me. I am collecting Lat/Lon information for the
> SC/FIA demo topology. So this should be available in the FIA/SC
> topology via the hasLocation relations from the Network resource.
>
> When the Automated Earth app does a detailed Query() it should discover
> the data path. Then, for each network in the path, it can retrieve the
> hasLocation resource from the topology. Thus it will have the
> hop-by-hop geolocation.
>
> I suggest we try a test run. If you have some code, perhaps Henrik could
> jam it into the Open NSA to throw a state change event at the
> AutomatedEarth app.
>
> I should have the initial topology diagram available by tomorrow for the
> automated GOLE call.
>
> This will be *so* cool. (:-)
> Thanks!
> Jerry
>
> On 10/5/11 10:45 AM, Tomohiro Kudoh wrote:
> > Hi Jeroen, Jerry and all,
> >
> > I propose the attached architecture for visualization from the view
> > point of the "NSI service plane" (not data plane like pinger).
> >
> > KDDI (Ikeda-san) may be able to configure their google earth based
> > visualization tool to use this architecture, and AIST is able to
> > prepare another more schematic visualization tool.
> >
> > To realize this, each ultimate provider should push state transition
> > information to the google app engine. If everybody agree, we will
> > prepare a sample code which should be incorporated to the
> > implementations.
> >
> > How do you think?
> >
> > Tomohiro
> >
> > 2011/9/28 Jeroen van der Ham<vdham at uva.nl>:
> >> Hello,
> >>
> >> My intention with the Dtox schema was to create something that help get the Automated GOLE demo started. The current topology mostly models control-plane aspects of a network topology, that is NSAs, NSNetworks and STPs (network edge ports).
> >> At some point these objects have to be mapped to a data-plane, and it was my impression that at SC we will start with that. At that point NML should come into play.
> >>
> >> There are already some NML related concepts in the auto-gole.appspot.com editor, but these have not been used yet. They have been simplified a bit, for example we're using bi-directional ports at the moment still.
> >>
> >> I do hope to apply more of NML topology to this, and hopefully this is a good opportunity to do that some more.
> >>
> >> Jeroen.
> >>
> >>
> >> On 27 Sep 2011, at 18:47, Jerry Sobieski wrote:
> >>
> >>> Hey Jeff-
> >>>
> >>> The NSA codes being developed by the various organizations are
> >>> definitely intended for a real system. But there is a lot of work yet
> >>> for these groups to fully implement the NSI framework and the CS
> >>> protocol. And since there is so little time between now and SC we are
> >>> trying to minimize changes the interop demo that would distract the
> >>> teams from writing NSI code. ...primarily the main change will be to get
> >>> the NSI agents covering real hardware present in the AutomatedGOLE
> >>> fabric. Hopefully, this will be fairly simple.
> >>>
> >>> The NML Topology is not adopted as the NSI Topology at this point. This
> >>> is an issue for the working group and won't be decided on the way to a
> >>> demo. Currently, we are using some extensions to the DToX ontology and
> >>> an RDF/OWL data representation of the topology for common presentation.
> >>> These seem to be performing well, but it is not clear to me how closely
> >>> this parallels NML. There are a lot of issues that need to be
> >>> discussed and understood regarding the NSI framework and how it uses
> >>> topology knowledge. The NSI Topology is on the working group agenda
> >>> for ongoing discussion (along with other topics) in parallel with the
> >>> demo effort.
> >>>
> >>> The Observer tool we were discussing for the demo could be a long term
> >>> tool. It would need to be able to Query() an NSA for the service tree
> >>> information down to the leaf nodes given a set of keys (e.g. Circuit
> >>> ID). This would essentially provide a monitoring agent with all the
> >>> information it needs to trace the circuit and know its state. The
> >>> interface to the NSAs could currently be done using NSI Query requests
> >>> in a polling fashion. I know polling sucks, but it would work for the
> >>> demo. Later on, or as a quick and dirty hack for the demo, a
> >>> notification system might be slipped in to replace the polling.. But
> >>> the high level logic would remain the same: The purpose of the observer
> >>> is to present the relevant information regarding a connection to the
> >>> user (or the ops/eng personnel, etc.) The details of the information
> >>> and how such information might be organized and visually presented is
> >>> the real value of the Observer. The Observer would also be able to
> >>> trigger other activities when conditions warrant. So I do not think
> >>> developing an Observer agent/application would be a simple throwaway
> >>> demo-ware exercise. Indeed, I think it would help us resolve formal
> >>> issues regarding external access to the service tree.
> >>>
> >>> I hope this helps. If you want to talk more details let me know and we
> >>> can put a conf call together with Radek to put more meat on the bones.
> >>> We have interest from some others as well.
> >>>
> >>> Thanks
> >>> Jerry
> >>>
> >>>
> >>> On 9/26/11 1:12 PM, Jeff W. Boote wrote:
> >>>> Will the OGF NML topology format be used? Or is this a temporary situation again?
> >>>>
> >>>> If this is something being built towards an eventual real system, I may be interested in putting resources into it. If this is a one-off for a demonstration, then probably not.
> >>>>
> >>>> jeff
> >>>>
> >>>> On Sep 22, 2011, at 8:06 AM, Radek Krzywania wrote:
> >>>>
> >>>>> Hi all,
> >>>>> For SC demonstration there is a need to have a centralised tool which will be able to show global reservation state and visualise the NSI demo topology. The first draft of such application will be required in about 4 weeks for Future Internet Week event held in Poznan between 24-28 Oct, where pre-SC NSI demo will be shown.
> >>>>> During my discussion with Jerry Sobieski, we would like to propose the following specification for this visualisation tool:
> >>>>> - A simple WebService needs to be implemented at network managers, which will allow to register event listeners (a WS addresses in fact), which will be the visualisation tool
> >>>>> - NSI enabled network managers will notify listeners on all events regarding reservations (mostly reservation status change according to NSI reservation lifecycle model)
> >>>>> - A visualisation tool will have a WS interface to receive the notifications from all domains involved in the demo
> >>>>> - A visualisation tool will be able to show global status on the screen, possibly a topology map. It would be good if a reservation paths could be shown (either delivered by NSI implementations or "guessed" by a tool having knowledge on involved domains)
> >>>>>
> >>>>> The way of showing a view is a subject to discuss, it can be either a web page or an standalone application. The details should be clarified soon, at the beginning of the next week.
> >>>>>
> >>>>> As developers who are already involved in coding the NSI and their network managers, it will be best to find someone else dedicated especially for visualiser task. Jerry and I will appreciate any volunteering developers here, so please if you feel being able to write a simple tool to support with a little graphical interface for NSI demos, please rise your hand. In case of any questions, feel free to ask on mailing list or directly to me or Jerry.
> >>>>>
> >>>>> Best regards
> >>>>> Radek
> >>>>>
> >>>>> ________________________________________________________________________
> >>>>> Radoslaw Krzywania Network Research and Development
> >>>>> Poznan Supercomputing and
> >>>>> radek.krzywania at man.poznan.pl Networking Center
> >>>>> +48 61 850 25 26 http://www.man.poznan.pl
> >>>>> ________________________________________________________________________
> >>>>>
> >>>>>
> >>>>>
> >>> _______________________________________________
> >>> nsi-wg mailing list
> >>> nsi-wg at ogf.org
> >>> http://www.ogf.org/mailman/listinfo/nsi-wg
> >>>
> >>>
> >>> --
> >>>
> >> _______________________________________________
> >> nsi-wg mailing list
> >> nsi-wg at ogf.org
> >> http://www.ogf.org/mailman/listinfo/nsi-wg
> >>
--
Takatoshi Ikeda <ikeda at ote.kddi.com> *** Changed E-mail address ***
More information about the nsi-wg
mailing list