perfsonar development work

["Matthias K. Hamm" <>]

Hi Maxim,

please excuse the late reply, I have lost your mail somewhere in my 
folders...

maxim schrieb:
> What is the difference between ENOC and E2ECU ?I thought the title E2ECU 
> supposed to mean that it monitors status of the End-to-end links and
> in case of monitoring limited only  by layer 2 there is no way to say
> that end-to-end path up or down for sure. And with LHCOPN cuircits its very
> important to be sure if the whole link is up or having some kind of
> problems.
>   
The Geant2 E2ECU concentrates on L2 because the service Geant2 provides 
is L2-connectivity which users can build an own OPN upon.
The E2Emon system was designed for supporting the needs of the E2ECU.

But I fully agree with you. We definitely need monitoring systems that 
can combine L2 and L3 metrics to cover real end-to-end measurements 
(because incidents don't care about organisational boundaries ;-) ) This 
means not nececarily need adding metrics to a monitoring system - it 
could also mean to combine existing systems (e.g. in perfSONAR we have 
quite a bunch).

In the last Geant2 Technical Workshop we talked a lot about combining 
monitoring systems to enable the things you pointed out.

Perhaps we could talk in the next LHC OPN meeting in Utrecht (together 
with the perfSONAR people and others interested in monitoring issues), 
exchange ideas and perhaps think about a roadmap ?

Cheers,

M&M


> --Maxim  
>
>   
> > -----Original Message-----
> > From: Matthias K. Hamm [mailto:] 
> > Sent: Friday, June 01, 2007 8:49 AM
> > To: maxim
> > Cc: 
> > ;
> >  'Mark Yampolskiy'; 'Phil DeMar'
> > Subject: Re: [pS-dev] Integration of E2Emon MP in next 
> > perfSONAR release
> >
> > Hi maxim,
> >
> > we agree that proper monitoring strategie always has to rely 
> > on different monitoring techniques.
> >
> > IMHO it makes no sense to put it all in one service. In our 
> > case, the ENOC may be interested in layer 3 measurements - 
> > the E2ECU is not.
> >
> > As perfSONAR is a service-oriented multi-domain monitoring 
> > architecture, it should be aimed to combine the measurements 
> > of multiple services (L2, L3, pingER ...) to get a "holistic" 
> > view - e.g. in CNM or other visualization tools.
> >
> > That should be adressed in the perfSONAR community, as it is 
> > a more general issue and covers not only E2Emon.
> >
> > What do you think?
> >
> > Cheers,
> >
> > M&M
> >
> > maxim schrieb:
> >     
> > > Hi Matthias,
> > > We've had a  discussion recently with Joe Metzger about how E2E 
> > > monitoring system is treating the status response from intermediate 
> > > MPs and from the endpoint MP.
> > > Currently they treated the same way and it makes the whole 
> > >       
> > monitoring 
> >     
> > > is kind of limited, since endpoint of the link has much bigger 
> > > capabilities in monitoring. For example the virtual circuit 
> > >       
> > from CERN 
> >     
> > > to FNAL could be down due BGP problems or any other routing 
> > >       
> > issues but 
> >     
> > > the layer 2 status will be up on every sub-link. Have you 
> > >       
> > though about 
> >     
> > > adding another E2E Layer3 MP into the LHCOPN monitoring 
> > >       
> > framework or 
> >     
> > > any other monitoring system or adding Layer 3 monitoring 
> > >       
> > capabilities 
> >     
> > > into the current E2Emon MP?
> > > Thanks,
> > >          Maxim.
> > >
> > >   
> > >       
> > > > -----Original Message-----
> > > > From: Matthias K. Hamm 
> > > > [mailto:]
> > > > Sent: Thursday, May 24, 2007 7:06 AM
> > > > To: Loukik Kudarimoti
> > > > Cc: 
> > > > ;
> > > >  Mark Yampolskiy
> > > > Subject: [pS-dev] Integration of E2Emon MP in next 
> > > >         
> > perfSONAR release
> >     
> > > > Hi Loukik,
> > > >
> > > > we seriously reworked our E2Emon MP to handle 
> > > >         
> > document/literal SOAP 
> >     
> > > > calls. We would like you to initiate the retesting in order to get 
> > > > the MP included in - hoepfully
> > > > - the very next perfSONAR release.
> > > >
> > > > Before going into integration tests done by the testing 
> > > >         
> > team, could 
> >     
> > > > you please give us more information about the testing environment 
> > > > (which
> > > > client(s) you test with, which request etc.)? We then do some more 
> > > > internal tests and hand you over a new RC then.
> > > >
> > > > See you soon in Cambridge,
> > > >
> > > > M&M
> > > >
> > > >
> > > > --
> > > >
> > > > Matthias Hamm
> > > >
> > > > ------------------------------------------------
> > > > Leibniz-Rechenzentrum / Leibniz Computing Centre Raum I.2.107 
> > > > Boltzmannstrasse 1, 85748 Garching, Germany
> > > > Telefon: +49 89 35831-8832
> > > > Fax:     +49 89 35831-9700
> > > > E-Mail:  
> > > >
> > > > -------------------------------------------------
> > > >
> > > >
> > > >     
> > > >         
> > >   
> > >       
> > --
> >
> > Matthias Hamm
> >
> > ------------------------------------------------
> > Leibniz-Rechenzentrum / Leibniz Computing Centre Raum I.2.107 
> > Boltzmannstrasse 1, 85748 Garching, Germany
> > Telefon: +49 89 35831-8832
> > Fax:     +49 89 35831-9700
> > E-Mail:  
> >
> > -------------------------------------------------
> >
> >
> >     
>
>
>   


--

Matthias Hamm

------------------------------------------------
Leibniz-Rechenzentrum / Leibniz Computing Centre
Raum I.2.107
Boltzmannstrasse 1, 85748 Garching, Germany
Telefon: +49 89 35831-8832
Fax:     +49 89 35831-9700
E-Mail:  

-------------------------------------------------