| Sebastien Goasguen is an Assistant Professor in the School of Computing at Clemson University who focuses on the design, modeling and provisioning of grid infrastructures for Virtual Organizations. He teaches undergraduate and graduate courses in distributed and parallel computing. But Sebastien also puts his teaching into practice and has worked with Clemson's central IT organization to deploy Condor. With Condor deployed, the Clemson campus grid was born and has now served over 2M hours in 6 months -- hours used by the faculty in industrial engineering, economics, chemistry and other departments.
Sebastien first encountered the OSG while working at Purdue University, where he teamed up with Norbert Neumeister on the Purdue CMS Tier-2 site. Here, he headed the deployment project of Condor on HPC resources for its use in both the OSG and TeraGrid. As the Purdue site lead, he pushed for the use of Condor resources on TeraGrid and for the interoperability of OSG and TG middleware.
Currently, Sebastien works with the OSG engagement team from RENCI. He is ensuring that the Clemson campus grid may be used by the OSG at large, while providing Clemson faculty use of other OSG resources in turn. Thus, the true spirit of grids and resource sharing is alive and well at Clemson. Next, Sebastien will use his operational experience of OSG resources to help other campuses build a campus grid and join the OSG.
|
| The fog has lifted. A new clarity is coming to OSG's collaborative workspace, the wiki. It will of course continue to evolve; that is its nature. The wiki home page and navigation bars now provide a clearer distinction between the different wiki webs and their purposes.
The Documentation web serves as central hub of OSG community-developed documentation. It guides the reader through the “whys and wherefores” of OSG and provides context-appropriate links into the release-specific installation and configuration pages.
The 0.8.0 release documentation now has a workbook-like structure, and includes some really useful new pages, for instance “Site Planning” and “Why upgrade?” Older, still really useful pages have been updated, in some cases reorganized, and importantly, reviewed by content experts. Archived documentation from previous releases is available.
Of course, we (the editors and reviewers) think it's clearer. The true test is YOU. Please don't hesitate to email osg-docs@opensciencegrid.org with any suggestions for letting more sun in.
~ Rob Gardner, Alain Roy & Anne Heavey
|
  |
|
|
Don Petravick |
As you know, Don Petravick has been the Security Officer for the Open Science Grid for nearly two years, during which time he has used his considerable expertise and vision to establish and communicate the foundations of our security principles in our open scientific environment. Through his leadership and example, Don has taught us how to identify
vulnerabilities and threats and to assess the risks that they bring. Based on his broad knowledge, Don established our Security Plan, bringing technical, operational and managerial activities for a secure infrastructure and program for resources, services and users.
On December 1 st, Don is stepping down from his post to concentrate on his many other responsibilities as Associate Head of the Fermilab Computing Division for Scientific Facilities. We will all miss Don's intellectual contributions, but we look forward to his continued involvement to OSG and will be calling on him in many areas. First and foremost, Don will continue to mentor our new Security Officer, Mine Altunay. He will also contribute to the OSG blueprint and technical roadmap. His responsibilities towards the network infrastructure on which OSG depends will also continue.
Mine Altunay |
I would like to take this opportunity to introduce Mine as she steps into this role. Mine has a PhD in computer science from North Carolina State and is part of the Computer Security Department in Core Scientific Services quadrant at Fermilab. She has ramped up quickly since she started with OSG in July, and devotes full time to OSG activities. I see her relying on the considerable experience and efforts of Doug Olson, Bob Cowles and the rest of the OSG Security Team, as well as the whole of OSG as she goes about her responsibilities.
Many thanks to Don! Welcome Mine and good luck!
~ Ruth Pordes |
| Around twenty OSG consortium members had booths at this year's Supercomputing conference. Below are summaries from a couple participants.
OSG - Education: The Education program at SC'07 was represented via tutorials, a poster and booth presentations. Our tutorials -- similar to our regular 3-day grid schools, but compressed into a one- day version -- provided an introduction to grid computing, giving the participants the opportunity to gain hands-on experience with globus and condor. They were very well received, with an audience consisting mainly of college faculty in disciplines such as chemistry, physics and computer science. The poster and booth session engaged participants as well, and.we noticed an increased interest in our programs, especially I2U2, our effort targeting high schools. We established some contacts which we will follow up on to develop collaborations and increase usage of the OSG resources in high schools and universities across the country.
For more information, including materials presented at the conference, see opensciencegrid.org/Education
~ Alina Bejan, OSG Education Coordinator
US-CMS: We demoed an OSG application running a CMS user analysis on the grid. Twelve slots were reserved from the University of Nebraska in order to shorten the job startup time, which led to jobs starting immediately. Of course, a lot of general discussion took place about why the LHC and CMS are exciting and why high energy physics researchers need to use the grid for data and processing requirements. According to Gutsche and Vaandering, there were several interesting people visiting the booth:
A woman from PBS Pro who is currently working on the admin documentation of pbs was interested in the technical details on how to administer pbs. With the demo, we could actually show her PBS Pro in action in the grid world because Nebraska uses pbs.
Yukiko Sekine, Ph.D., Program Manager for the DOE Office of Advanced Scientific Computing Research was interested in analysis submission and how CMS developed and could improve it's Grid work flows.
A visitor from Advanced Micro Devices complained about recent BBC science programs being shallow and not demanding enough. He was happy to talk to the scientists about the scientific prospects of grid use in terms of Higgs and Supersymmetry.
~ Oliver Gutsche & Eric Vaandering, Fermilab
|
|
|
 Britta Daudert |
A two-day focused meeting was held recently with the goal to develop a concrete six-month roadmap for having LIGO Inspiral work-flows running in production across multiple LIGO Data Grid (LDG) clusters as well as other resources, including those of the OSG. Participants included members of the LIGO Inspiral group, the Condor group, the Pegasus group at ISI and me, a liaison between LIGO and the OSG.
First, each group contributed to a series of talks giving a general overview of how the analyses are structured, what the capabilities of the technologies are, what has been accomplished so far and what problems were encountered along the way. My talk, “Inspiral efforts on OSG: the good, the bad, the ugly” focused on showing the Inspiral scientists that it is relatively easy to run LIGO work-flows on the OSG using the Pegasus work-flow planner developed at ISI. In fact, it's not much different from submitting work-flows to the LDG via Condor DAGMan, as it is currently done. Using the Pegasus planner has the benefit of explicitly managing data in the work-flows. With the recent joint work between USC and UW Madison on an end-to-end work-flow management system, there is an opportunity to leverage both technologies for LIGO analysis.
Group discussions followed, identifying missing pieces and beginning talk about realistic solutions for a successful collaboration of the groups involved in the project. Most of day two was used to solidify the six-month plan by developing detailed tasks to be achieved by each group.
This workshop was very different from any other conference or workshop I have attended over the years in that actual work towards the common goal of the meeting was done. Together with members of the Inspiral group, a member from Pegasus and a Condor team member, we broke off into a different room and spent the morning and part of the afternoon teaching each other what we knew, writing a shell script together and voila! By the end of the meeting, we were successfully submitting work-flows to the LDG with the Pegasus work-flow planner.
Needless to say, everybody was impressed at how well we worked together and how much we achieved as a team, sharing our knowledge rather than struggling alone in our separate groups. I believe that all of us walked away feeling elated by the success of this workshop. It just goes to show, all it takes is a small group of motivated people with the right attitude and willingness to share and work together.
~ Britta Daudert, Caltech
|
|
