 |
|
DAILY NEWS AND INFORMATION
FOR THE GLOBAL GRID COMMUNITY /
|
Special Features:
CHARLIE CATLETT SPEAKS ABOUT
LEAVING GGF
By Alan Beck, Editor-in-Chief
Last week at GGF10 in Berlin, Charlie Catlett, the founding chair of the
Global Grid Forum, announced that he would be leaving after nearly five years
in the position. In this interview he discusses future plans for the GGF and
himself, and comments on the positive and negative aspects of competition in
the standards community.
ALAN BECK: How will your departure affect the GGF? Should
your
office
be
filled by an individual with an academic, governmental or commercial
background, or should other considerations be the determining factors in
selection?
CHARLIE CATLETT: Well, first of all I'm not leaving GGF.
I'm simply stepping aside to make room for a new GGF chair after nearly five
years. Today, GGF is led by a dedicated steering group with the help of a very
strong secretariat led by Steve Crumb, our Executive Director. I'll continue
to be part of the steering group and will be participating in GGF while also
encouraging more people from deployment projects to jump into GGF.
Where we want the next GGF chair to focus is on strategy, collaboration
with
other standards groups and communication -- internally and externally.
In terms of strategy, we have a growing body of work centered around our
Open
Grid Services Architecture (OGSA), which we have developed from a series of
use cases to support enterprise or "utility" computing and information
applications. This year we have begun to capitalize on the OGSA roadmap as an
organizing principle to the work of GGF. Our collaboration with other
standards groups is also a top priority. Last year we began to build
relationships with other standards bodies, and this continues to be an
important focus. Currently we are doing a lot of collaboration with OASIS
because of the fact that OGSA is layered on top of Web services. Finally, as
any organization grows and becomes more visible it is increasingly important
to communicate consistently what you are about, both internally and
externally.
We are really looking for someone who can take GGF forward with a strong
vision, and we are not focusing on whether they should be from a commercial or
academic background. It has to be the right person, and that person has to be
able to lead a community that is both commercial and academic.
AB: Please define the process by which your successor will
be chosen. How can candidates present themselves for consideration?
CC: Well, GGF is managed by a community-selected steering
group of eight leaders from industry and twelve from universities or
laboratories. The steering group will serve as the search committee for the
new chair, working closely with our external advisory committee which is
chaired by Paul Messina. Dennis Gannon (Indiana University), Ian Baird
(Platform Computing), Ian Foster (University of Chicago and Argonne National
Laboratory) and Peter Clarke (University College London) are leading the
search effort. We are asking people to nominate candidates to any of these
individuals.
AB: What are your general observations about GGF and its
role? How do you see these evolving over the next few years? What major
challenges will present themselves?
CC: GGF is filling a very important role that goes well
beyond just doing Grid standards. We bring end users together with software
developers, making sure that standards are really driven by requirements. As
importantly, we bring together industry and academic leaders to ensure that
our work is commercially relevant and viable while also bringing in the best
ideas from the research sector into the marketplace. The community, or
"forum," aspect of GGF is something that we have built over five years and
it's one of our real strengths. I recently looked at the participation in the
OGSA work, for instance, and found that 56 companies were involved along with
40 other organizations -- in all over 500 individuals. Getting agreement at
that scale requires trust and community.
When we started GGF, we really focused on building the community and
getting
people to work together focusing on specific problems and requirements. Right
at this moment the GGF standards work is just starting to heat up. We are
heavily involved in helping to define the shape of Web services, and we now
have a roadmap -- OGSA -- that provides a lens through which we can see
specific gaps in Grid technology that need to be filled in. Coupled with this
we have dozens of high-quality documents and specifications that are working
their way through the standards process. So I think one of the things that
will characterize GGF over the next few years is the quality of the standards
in the document series.
One of the challenges we face is that Grid technology is becoming very
popular, and there is a huge amount of hype. We have to be realistic about the
state of the technology and work with the consumers of that technology to move
it forward rather than get too far ahead of them guessing what they might want
or need. Related to this popularity, GGF has been characterized as an academic
organization looking primarily at science applications, and this is an
outdated view of GGF. We are a community of over 80 companies as well as
universities and laboratories from 40 countries, and we are focusing very
heavily on commercial applications. The use cases driving the OGSA roadmap
were not physics and chemistry or supercomputers -- they were commercial
applications and use cases coming from companies.
AB: Should other entities or bodies "compete" with the
GGF? Why or why not?
CC: We've not encountered a competitive spirit in the
standards community- it's very cooperative in our experience. I think this is
because people realize that it's not productive to have competing standards
bodies -- there is too much work to do, and the best approach is to
collaborate. At the same time, it's important to respect the work that is
happening elsewhere, where there is momentum and coherence. A recent example
of this mindset in GGF has been the WSRF and WSN work, which grew out of the
GGF OGSI work. We fully support this new Web services work happening in OASIS,
because that's where Web services are converging.
When we started out five years ago, we were mentored by leaders from the
Internet Engineering Task Force (IETF). In one of our first conversations with
the IETF we proposed that GGF join the IETF as an area focusing on middleware.
For a variety of reasons, including their own explosive growth, the IETF
recommended that we continue to create a compatible but new organization to do
middleware work. So, even in our own case, we first looked at trying to do our
work in an existing body and only after talking to them did we decide to form
a new group. Today, we continue to enjoy good collaboration with the IETF.
We partnered with the Distributed Management Task Force (DMTF) last year
and
this year we are working closely with OASIS. We've started discussions with
W3C and have an open door to pursue that relationship as well. In each case,
we look at what we are doing and how we can contribute to the overall body of
standards, and our particular strengths lie in developing Grid "solutions"
that guide the application of lower-level standards such as specifications
from GGF, Web services from OASIS or W3C, or Internet standards from IETF.
One might look at a snapshot of GGF today or at some point and say "well,
you
are not focusing on this or that" but the GGF process is such that when a
group of people comes together and says "we want to work on this" we can
create a working group or research group within about two weeks. So, if we are
not working on a particular technology or use case, it's not because it's
necessarily out of scope or off the radar screen for GGF, but because we have
not found the group of consumers or producers who are energized to work on
that problem. That's the point of an open process.
It's good to have competing approaches when it's not clear what is the best
way to solve a technical problem. I think it's not good to have competing
activities driving or developing standards, because it confuses the market and
dilutes the value of the work.
AB: What are your own plans for the future?
CC: Well, I've been working on the TeraGrid project which
is funded by the National Science Foundation and now is an important time for
me to focus my full attention on that project. This year we are in the process
of expanding from five to nine sites. We built the network and we are just
about finished deploying the computing and storage resources -- so now the
real fun begins from the point of view of Grid technology and interacting with
end users of that technology. I'm working closely with Pete Beckman to create
a software, integration and coordination capability based at the University of
Chicago and Argonne National Laboratory, partnering with the other TeraGrid
sites to develop a plan for evolving the TeraGrid over the next five years.
The TeraGrid is also a growing community, and I'm eager to get more involved
in deploying a real grid system with real users.
|