This page is old; please merge its content into the Wiki and delete it!
ROSCOE_20091002
ROSCOE Phone Meeting (2 October 2009)
Actions and important decisions taken during the meeting.
Budget targets sent earlier in the week
The targets were:
MGT: 500 kEUR
HEP: 2500 kEUR
LS: 1000 kEUR
CCMST: 1000 kEUR
GO: 1000 kEUR
CS: 750 kEUR
PS: 750 kEUR
H: 500 kEUR
These limits were generally accepted except for HEP which will need an envelope of 2700 kEUR and PS which needs an envelope of 1000 kEUR. The envelope for PS will be looked at more carefully once the detailed work plan following the new activity layout is available. So the current targets are the following:
MGT: 500 kEUR
HEP: 2700 kEUR
LS: 1000 kEUR
CCMST: 1000 kEUR
GO: 1000 kEUR
CS: 750 kEUR
PS: 1000 kEUR
H: 500 kEUR
These targets neglect direct costs for dissemination (probably around 50 kEUR) and funds for inviting people to the EGI User Conferences (also probably around 50 kEUR). It also neglects costs for the overall coordination of dissemination/training which I'd estimate to be around 15 PM of effort.
Given the tight budget, ROSCOE will continue to aim towards a 3 year project. All task and effort planning should use 3 years as the basis.
Reorganization of the project activities
To date the project has been organized by SSC. This organization leads to a project with around 30 separate activities--a number which is probably not manageable. Including only one activity per SSC is not really possible because the EC requires that the tasks be separated by type (networking activity, service activity, and joint research activity).
Gabriel did a nice analysis of the current work plans and proposed a activity reorganization based on common tasks. Cal has taken that input and produced a new "combined" draft that generally follows that reorganization (combined draft V4 available from trac). This reorganization is:
NA1: Project Management
NA2: SSC Coordination
NA3: Dissemination & Training
SA1: User Support
SA2: Scientific Gateways
SA3: Targeted Application Porting
JRA1: Data Management Evolution
JRA2: Quality and Credit System
JRA3: Contrib. to sustainable ops. model; Grid ontology
JRA4: Remote Facilities and Interactive Grid
The draft also contains references to tasks in the existing work plans that fall into the new work packages. To my knowledge, all of the existing tasks are incorporated into the new organization.
It is important to keep an "SSC Identity" even though the work plan is now organized thematically rather than by SSC. Because of that, the tasks NA2, NA3, SA1, SA2, and SA3--all have one subtask per SSC. The SSCs are expected to manage those subtasks. In addition, the project management will still be done by SSC Coordinators and NOT by activity managers.
Comments on individual activities
NA1
No comments.
NA2
It will be important to identify what specific international projects will be contacted for liaison with each SSC. The individual SSC descriptions should include this detail. In addition, activities such as VO coordination, resource allocation, and monitoring of use should be included here.
NA3
The training description should be expanded and it should be made more clear that the training effort involved in the SSC is specialized expertise that is added to the standard, generic training. In addition, the individual SSC descriptions should contain information on important events or training targets for the SSC.
SA1
No comments.
SA2
There was a confusion as to where workflows would be handled. Workflows will be handled in the scientific gateway work package; the SA3 description should change "workflows" to something like "standard use cases".
SA3
This is an area that can have large impacts with a relatively small investment in manpower. Each SSC will have to have a policy on how applications/software is chosen for porting. For political reasons, we'll probably not want to have a complete list of applications/software to port in the proposal. This activity can also provide a bridge with PRACE as many of the scientists will want to use both DCI and supercomputing resources. It would be good to add a common task to survey what software is available elsewhere.
For JRAs see the next topic.
Concrete synergies/collaborations between ROSCOE SSCs
ROSCOE must demonstrate that there is more value in having the ROSCOE SSCs in a single project than having seven different SSC projects. That is, we must show that there are concrete synergies between the SSCs. We can do this either by adding common subtasks to the NA and SA activities and/or creating JRAs that span more than one SSC.
The current JRAs in the Draft V4 were just what was "leftover" from the NAs and SAs in the current scheme. They are:
JRA1 Data Management Evolution
This is still an important topic for PS. It was dropped from the HEP program of work because of the budget constraints.
JRA2 Quality and Credit System
This was defined by the CCMST SSC, but they would like to see others interested and participating in this.
JRA3 Contrib. to sustainable ops; Grid ontology
Defined by the GO. HEP has suggested that it may make sense for them to contribute here in terms of sustainable operations.
JRA4 Remote Facilities and Interactive Grid
Defined by PS. They are currently in contact with some other projects focused on these objectives and this task may disappear from ROSCOE.
In general, concrete suggestions for common, synergistic tasks or ways to define the JRAs more broadly to have a larger participation are welcome.
Meeting with EU Project Officer
Cal will be meeting with the Project Officer for the Virtual Research Communities call on 8 October. He will circulate a set of questions that he intends to ask early next week. If others have questions they'd like asked please send them to Cal ASAP.
Schedule for having partner information
The deadline for having A2 information for all partners is 23 October. Coordinators are asked to push their partners to provide this information ASAP and no later than the deadline. This must be done as early as possible to ensure it does not block the proposal submission. After all of the A2 information is available, then work on the detailed budget can begin (assuming that the detailed work plan is also available).
Schedule for completing proposal
The goal is to have a "complete" proposal by the end of October. This means that all of the sections have been completed and are of reasonable quality. Detailed work on the budget and fine-tuning of the proposal will take place during November.
The schedule will be the following:
9 October
- SSC coordinators provide descriptions of subtasks for all
NA and SA work packages including the effort allocated by partner. (Sooner if possible.)
- Cal will provide outlines for all of the remaining sections of the
proposal.
Volunteer to "reduce" impacts section found.
Suggestions for common subtasks for work packages or combined JRA work
packages received.
16 October
- SSC coordinators will provide dissemination, exploitation plans for
their SSCs. They will also provide list of socio-economic impacts.
Impacts section combined and rewritten.
Volunteers found for combining dissemination, exploitation, socio-economic impacts sections.
30 October
All parts of the proposal available including partner contributions.
The goal will be to provide a new combined draft after each of these
deadlines. People working on the draft should use the previous combined draft and make changes (with change tracking on!) on that draft. Cal will combine the changes in the next combined draft.
AOB
Letters of Support
ROSCOE will ask for letters of support from:
EGEE-III
EGI
WLCG
ESFRI projects affected by ROSCOE
Centre for Autonomic Computing (CAC)
CoreGrid
SSC coordinators are asked to identify ASAP which ESFRI projects should be approached for a letter of support. Cal will work this coming week on sending out the requests for the letters of support. Please inform Cal of any additional projects that should be contacted ASAP.
Face-to-Face Meeting
Antonio has another meeting the day before and wondered if others have such travel complications. In any case, if travel to the face-to-face meeting is easier if the meeting time is shifted a bit, please let Cal know as far in advance of the meeting as possible.