NA4 Steering Committee Meeting
CERN
NA4 SC Phone Call
November 26 - 14:30 CET<o:p></o:p>
Participants:<o:p></o:p>
- Vangelis (chair)<o:p></o:p>
- Mariusz<o:p></o:p>
- Johan<o:p></o:p>
- Christos<o:p></o:p>
- Cecile<o:p></o:p>
- Christophe<o:p></o:p>
- Horst<o:p></o:p>
<o:p> </o:p>Agenda topics:<o:p></o:p>
<o:p> </o:p>a. Quarterly report 6<o:p></o:p>
<o:p> </o:p>Kristina requested input from certain teams. So far I’ve seen response from:<o:p></o:p>
<o:p> </o:p>HEP (Patricia - Is this OK? I didn’t see any attachement in the mail)<o:p></o:p>
LS (Johan send a list of publications)<o:p></o:p>
DUS (Christos replied to Kris, I haven’t seen the input though)<o:p></o:p>
APS? Apart from the specific request for APS Gergely could also provide feedback for the seed resources metrics?<o:p></o:p>
Fusion?<o:p></o:p>
<o:p> </o:p>All: Any additional list of publications?<o:p></o:p>
<o:p> </o:p>Horst: We’ve send it directly to Kristina<o:p></o:p>
Cecile: We might have one more to add to the list.<o:p></o:p>
<o:p> </o:p>Vangelis: Any news about the Use Cases screencasts?<o:p></o:p>
<o:p> </o:p>Christos: Nothing more to report for the previous quarter. The AUTH team was experimenting with screencast tools this period and will deliver the first results in the coming quarter.<o:p></o:p>
<o:p> </o:p>Action to all: Check if you have any more publications to send, for the period Aug-Oct 2009, and send it to Kristina with CC to Cal and Vangelis<o:p></o:p>
<o:p> </o:p>
<o:p></o:p>b. Matlab/Lumerical licenses<o:p></o:p>
Matlab - Vangelis had meeting with Mathworks people at SC09. They haven’t progressed much with their interactions with EGEE sites (only with 2-3 of them). There has a been more closer interaction with the Belgium NGI. Working on a Utility licensing scheme, currently targeting Amazon EC2 but it might be useful also in the EGEE context. I’m currently investigating this.<o:p></o:p>
Cecile: To keep GO in the loop. Will give a technical person as a contact point.<o:p></o:p>
Lumerical: We’ve received a quite long list of interested users that would like to try out Lumerical. Does anybody recognize this people? Is there any specific interest from any of our clusters? <o:p></o:p>
The solution has already been tried on EGEE by the Scottish Grid:<o:p></o:p>
“We have worked with Lumerical and a local representative Jonathan Leach in order to deploy Lumerical's FDTD onto the EGEE Grid. As you probably know this is a licensed product and requires the use of a license server which the batch system will connect to in order to manage the licensing. In order to do this we use consumable resources passed to the batch system. This is hard to do with the lcg-CE but made much easier with Cream. As the job manager is much easier to enhance. The solution means you can actually send the resource request in JDL. Therefore, we support FDTD users on direct CREAM submission only. ScotGrid also has support for MPI jobs and we have wrapped up the the mpistart submission scripts into one easy to run FDTD script for local users to run. Using Cream also means that a gridftp server installation is required. Once you can submit to Cream via the WMS then the gridftp server on the WMS can be used for this purpose.” Douglas McNab, Scottish Grid<o:p></o:p>
Don’t know if the dependency on CREAM is strongly required or LCG-CE could be used equally without any problem.<o:p></o:p>
Next steps: Should find some resource centers that would be interested to deploy the software and offer it to users? Application Porting Support team could be assigned with this responsibility (maybe together with SA1 if we cannot allocate enough sites from our VOs). The seed resources could be also considered. <o:p></o:p>
c. GridwiseTech AdHoc<o:p></o:p>
The request from Pawel has come a few days after EGEE09 and it has been pending since then. AdHoc is a resource sharing platform that help users create ad hoc collaborations over the net and share data (mainly) and other resources.<o:p></o:p>
Johan was the only one to react with the following comments: <o:p></o:p>
“It seems to me that this product is presented (on paper at least) as a commercial alternative to the gLite stack. If the question is "will traditional users of EGEE be interested?" I think the answer is no as the strongly appealing property of the EGEE middleware for scientists is that it is open and distributed in many different centers.<o:p></o:p>
From what I see, there is no interface to gLite resources in the product nor clearly defined added value. Unless you were presented some specific plan for this tool.<o:p></o:p>
This tool is probably addressing communities who do not already have a grid solution and are looking for an accessible product to setup a grid with their resources.”<o:p></o:p>
Vangelis: I generally agree with the above point of view. I don’t there is a particular interest from our side for this platform. Nevertheless we could investigate potential interoperability with gLite. The insides of AdHoc’s architecture is not clear but from a first look it seems highly unlikely<o:p></o:p>
Action (Vangelis): Reply to Pawel requesting info more concerning the above. In principle there is small interest from our side if this concerns a solution that has no relationship with gLite.<o:p></o:p>
<o:p> </o:p>
d. UF5 update<o:p></o:p>
Progressing normally but slow. 5 submissions so far (two withdrawn, one of them was duplicated). The deadline is approaching so we better start getting more active by spreading the word and chasing our closest teams to prepare abstracts. <o:p></o:p>
<o:p> </o:p>Early bird registration is open. It was confirmed that after 1st of January the prices will be charged with 25% VAT thus it makes sense to register real early in order to avoid the extra charges. Spread the word among your people and be sure to register in the next coming days in order to give a good example<o:p></o:p>
<o:p> </o:p>Action to all: Start chasing your colleagues to prepare abstracts and also to registers as soon as possible.<o:p></o:p>
<o:p> </o:p>e. AOB<o:p></o:p>
<o:p> </o:p>None<o:p></o:p>
<o:p> </o:p>Meeting closed at 15:00 CET.<o:p></o:p>