Dear all,

I am opening the thread. Here are my thought on what could be NA4 in EGEE3.

The NA4 roles are:

 1. to steer middleware development
    (requirements, prioritization, testing new services)

 2. to help scientific communities exploiting the infrastructure
    (assistance, porting, demonstration of feasibility and induction)

 3. to make/demonstrate production use of the infrastructure
    (large scale usage, scientific production)


In EGEE2, these goals are difficult to reach for many reasons:

 1. Middleware steering.

    The requirement analysis as been done in details in EGEE phase 1
    already. There has been no addition to the requirements database
    in EGEE2 to my knowledge. However, triggering new middleware
    developments and making coordinated testing is something for which
    a formal process is not clearly established. This as been done on
    a case by case basis in the TCG working groups only. From my
    (biomed) experience, there is very little resources available for
    new development anyway thus strongly limiting this activity.

 2. Exploiting the infrastructure.

    There is a huge gap between partners involved in the project and
    benefiting from NA4 resources for their application development
    and external scientific communities with hardly any support at
    all. In the current model, NA4 resources are very widespread. It
    let little place for common interest actions (assistance and
    porting, operating some specific services). More over, there is a
    regular demand of external projects to get assistance from EGEE
    engineers to get started to which we cannot respond.

 3. production use.

    Achieved, especially by HEP, but we should probably ask for more
    feedback on scientific results.


Suggestions based on this analysis:

I think NA4 resources are too widespread. We should have a pool of
engineers (preferably located in a same place, just like JRA1 is doing
for efficiency) to work on common interest activities:

 - Operating application services (VOMS, File Catalogs, AMGA...)

 - Providing assistance to new comers

 - Dedicate effort to selected applications development (requires the
   selection of most interesting applications for grids, coming from
   internal or external partners, other EU projects, etc).

 - Common interest application-level services development (e.g.
   robust massive jobs submission as needed and developed concurrently
   by HEP, WISDOM and GROCK among others, short deadline jobs, etc).

In any case, we should definitely avoid to have fractions of FTEs
spread over tens of institutes for efficiency. This group would also
help in having close contact with JRA1 for representing application
needs and early testing of new services.

To sustain the effort invested so far, some FTEs should be allocated
de facto to successful application activities in EGEE2 but some
reasonable manpower should be left available for this common interest
activities.

We need to make a selection of applications in which to put our effort
based on criteria such as suitability of grids, added value and
capability to hold the target application. We should as well follow on
the selected project and get feedback on the scientific production to
rediscuss the affectation of manpower to different applications.

Regards,
Johan