WHAT ARE OGF EVENTS

THE NEXT EVENT

Registration

Registration Fees

Location/Local Info

Lodging

Schedule

Participants

Special Events

Chair Info

Student Scholarships

PAST EVENTS

GGF17

GGF16

GridWorld/GGF15

GGF14

GGF13

GGF12

GGF11

GGF10

GGF9

GGF8

GGF7

GGF6

GGF5

GGF4

GGF3

GGF2

GGF1

GF5

GF4

GF3

GF2

GF1

GF1 - The First Grid Forum
June 16 - 19, 1999
Moffett Field, California, USA
NASA Ames Research Center

The number of Grid initiatives combined with the overlap in people, technologies, and challenges of these early initiatives encouraged a number of Grid researchers and managers to convene the first "Grid Forum workshop", which was hosted by the NASA Numerical Aerospace Simulation (NAS) and in particular the Information Power Grid (IPG) program, held at the NASA Ames Research Center from June 16-18, 1999. Organizers expected perhaps 50 people, however attendance was over 150 from some 50 organizations and 4 countries, clearly indicating great interest in the concept. There was considerable enthusiasm for the establishment of a more permanent Grid Forum. The groundwork for that body, including the initial working group organization and discussion, was laid at this workshop.

Description

There is increasing interest in the use of networks not just for communication or remote data access, but also for the coupling of computers, instruments, data archives, etc., with each other and with humans. Some of these applications include collaborative environments, distributed computing, computer-enhanced instrumentation, distributed data mining, digital libraries, and problem solving environments.

Within the scientific community, this interest has led to new initiatives such as the NASA Information Power Grid, NCSA Alliance National Technology Grid, NPACI metasystems, DOE Collaboratories, DOE ASCI DISCOM, etc., as well as the establishment of focused community forums concerned with, for example, the use of Java for high-performance computing (Java Grande Forum), desktop access to remote resources (DATORR), component architectures, (CCA Forum), etc.

The Grid Forum is being proposed as a means of enhancing the effectiveness of these diverse activities by providing communication among practitioners and, in particular, the definition of standard infrastructure elements. For example, multiple groups (DOE2000, NSF PACIs, NASA IPG, and probably others) are working to define Grid information services used to discover and characterize computers, software, networks, etc.; there would seem to be considerable value to agreeing on standard data models and information infrastructure, and the Grid Forum can provide a forum for that standardization process.

The goals of this First Grid Forum Workshop are to:

• verify that there is indeed interest in proceeding with the  creation of a Grid Forum
• determine the "rules of engagement" under which a Grid Forum should operate
• identify and commence discussion on an initial set of topics

Wednesday, June 16

Welcome and introduction: Bill Feireisen (10 min)
Agenda and goals: Ian Foster (20 min)

Requirements and statements of interest

The meeting started with a series of presentations by representatives of Grid-related projects and activities. Each speaker was asked to provide a brief overview of the activity in question and to identify areas in which standardization and/or information exchange are desired.

• Information Power Grid: Bill Johnston
• Alliance National Machine Room: Randy Butler
• Alliance Access Grid: Rick Stevens
• NPACI metasystems: Andrew Grimshaw
• Data-intensive applications: Reagan Moore
• NSF workshop on online instrumentation: Klara Nahrstedt
• DISCOM: Judy Beiriger
• Internet 2: Ted Hanss
• DATORR: Gregor von Laszewski
• Java Grande Forum: Geoffrey Fox
• Alliance Portal Architecture: Dennis Gannon
• DOE2000 Collaboratories: TBA
• DOE Common Component Architecture: TBA
  

Candidate Grid Forum activities

Two broad-based activities have got started in the last year that appear to be candidates for Grid Forum sponsorship as well as a potential model for GF working groups. We heard from leaders of these efforts.

• Grid Security Infrastructure: Steve Tuecke (40 min)
• Grid Information Service: Brett Didier, Steve Fitzgerald, Karen Schuchardt, Gregor von Laszewski (40 min)

Thursday, June 17

In establishing a Grid Forum, we can look to a variety of models such as the Internet Society and Internet Engineering Task Force (IETF), the Message Passing Interface (MPI) Forum, etc. Choosing the correct "rules of engagement" for the Grid Forum will be vital to its success. Hence, we heard brief summaries of how various existing bodies work (IETF, MPI Forum, others), and then discussed a concrete proposal for the organization of the Grid Forum. This proposal featured an IETF-like model, in particular:

• As little formal structure as possible!
• A goal of advancing Grids as a whole by identifying, defining, and encouraging deployment of common infrastructure elements
• "Rough consensus and running code": i.e., develop standards based on experience with real implementations
• Focused, short-term standardization processes (e.g., what can be achieved in a year) rather than open-ended activities

Following this, we broke into discussion groups, with the goal of identifying topics within which initial Grid Forum activities should be contemplated.  The list of such groups will be determined according to where there is sufficient interest, as evidenced by someone putting an interest group and a draft list of discussion topics together prior to the meeting.  However, here are some candidates:

• Grid Security Infrastructure
• Grid Information Service
• Resource specification / resource reservation
• Collaborative toolkits - e.g. CIF
• "Global shell" functions
• Standardized scheduler interface
• Resource allocation management / accounting / accounting records
• Monitoring/auditing message/record functions and formats

Friday, June 18
Friday morning was spent in breakout groups, once more, plus a further plenary session.