Meeting of APPS-RG at GGF13, Seoul, Korea, March 15, 2005, 11:00-12:30
Notes taken by: Shantenu Jha, Thilo Kielmann
Agenda:
* 0:00 Welcome, opening
* 0:05 Status update
o new charter approved
o document GFD.39, "Workshop on Applications and Programming Tools"
has been published
o other documents in the pipeline
* 0:20 Report from SAGA-RG: progress on simple Grid APIs
* 0:30 Open microphone round:
"Current issues of grid application developing and deployment?"
* 1:15 Planning of future activities (e.g. workshops)
o GGF14 (with Earth Observation community???)
o GGF15 ???
* 1:30 Adjourn
Thilo Kielmann opens the session
GGF IPR policy is pointed out and sign-up sheet is circulated
Status update:
new charter has been approved, group name has been adjusted to
"application developers and users research group"
This raises the questionamong the audience:
- "What is an Application?"
Do we need a tighter definition? Kind of consensus in the audience:
- Application: Uses the Grid to get other things done.
Document writing:
GGF12 workshop report now published as GFD39.
We urgently need to initiate the other Workshop Reports
Will start with GGF13
Report from SAGA-RG: progress on simple Grid APIs
Andre Merzky gives a short presentation on the progress made within SAGA-RG
Open microphone round:
"Current issues of grid application developing and deployment?"
This is the so-called "Water Cooler Session" where people freely speak
up and tell what they are mostly concerend with (like at a company
water cooler)
The following statements try to catch what has been said. They are
personal statements, subject to inaccuracies of note taking.
Tomasz Haupt:
not all users are the same, has diverse needs
need infrastructure that is stable
have an interface that hides the changing stuff
i.e., similar to SAGA
Pascal Kleijer: quick implementation is critical
If SAGA can provide the interface
then use as a black-box will help
users don't want to change their code
therefore need non-mutating wrappers
(i.e. clean wrapper)
Grid Added value: seamless use of remote resources
Tom Goodale: "Currently don't find the grid useful at all"
lack of stability and consistency main problems
parallel environments: a stable one + a development
(like production and development)
Basic s/w and system adminstration paradigms not followed
Production Quality S/W Required.
Couple of first time users: (names unknown)
government & business users
in business important to have stable environment
Commonality b/w DRMAA and SAGA?
relationshp b/w such different API's?
common look and feel
NII users: interactivity & easier usage
KISTI user: how long will we have to wait for services like SAGA?
T.Goodale: 6-12 months.
Pascal K.: industry needs quick solutions
and are proceeding to have "solutions without standards"
A.Merzky: "solutions w/o standards" isn't Grid Computing
GGF Steering Group member: Need to strike a balance b/w
rapid development and deployment and standardization
2 "speeds" levels :
- Speed of deployment
- speed of integration of code & Usage
low level integration is most susceptible to change
as a user, fault tolerance is ususal hence Open Science
Grid build a few simple services which are Robust
and build others on top of it.
Suggestion: Demos are useful
Planning of future APPS-RG activities:
Production Grid Management:
APPS-RG is seen (by some) as the flagship of community building
APPS-RG has been approached for a "community panel" track for GGF14
as part of GGF's reorganization, user communities shall be strengthened
track at GGF14 will use existing groups and their results
question from the audience: APPS-RG's umbrella function?
(currently not addressed, for near future)
possible future workshop topics:
- Grid-Rel maybe a group to interact with
TK: prod grid testing, usability
(seems to be a 'hot topic' as mentioned in the waer cooler session)
will approach PGM-RG to discuss a possible joint workshop
meeting adjourned as the agenda had been handled completely