XSEDE Community Infrastructure

Campus Bridging - many means

2013-07-25T15:13:18Z

This week at XSEDE 13, Scott Michael presented an example of Campus Bridging activities with Mark Lynch's project making use of mlRHO. Scott and his collaborators identify a number of activities that helped the Lynch lab bridge from campus resources at Indiana University to XSEDE resources. These included:

* Indentifying that the team had big data needs, rather than HPC needs
* Requesting an allocation and creating a scalability study
* Adapting code to run with XSEDE Scheduler environments (they helped the team move from hundreds of small serial job submissions to small numbers of BigJob submissions)
* Profiling code to identify points for improvement - Lynch's team made adaptations to the code in order to carry those out

Scott's team was able, with not a lot of effort, to provide considerable improvement to the Lynch lab's analyses. Code improvements carried out by the user team resulted in a 50x speedup. Many of the activities involved improving the users' understanding of XSEDE processes and letting them know about technologies (BigJobs for submission, Vampir for code profiling) that helped them get the job done.

Campus bridging is in many instances the art of pointing the way to resources that already exist in a way that allows users to understand the benefits they can attain. I think there are a lot of opportunities for finding potential XSEDE users and projects that involve just this sort of connection that includes some advocacy, some documentation, information sharting, and some project-specific interactions with users.

Campus Bridging year 2 activities - Rocks rolls, use cases

2013-02-10T18:18:10Z

Hi all!

WE have a pretty well worked out strategy for distribution of cluster software. The basic idea is to make it easy for a cluster admin to stand up a cluster that mimics a "generic" xsede cluster. Worked out well and of good quality from the standpoint of intended users . . .those can be two things. So the current cut of that document is attached here. We are exspeically interested to hear from people on smaller campujses as to whether or not this approach seems helpful. A list of potential / planned software to be installed by default is included. We are especially interested in comments on that list.

Also attached here is the level 3 decomposition document for the campus bridging use cases.

RE: Campus Bridging Use Cases

2012-02-16T03:45:18Z

Wow, there is some really great stuff in here. Some of what is described applies to a much broader scope than I had expected. It's quite nice to see this detail so that we can have something with edges. I'm looking forward to close study. Nice!

Campus Bridging Use Cases

2012-03-06T15:21:28Z

The CB team, in collaboration with other XSEDE staff, have created some use cases for for Campus Bridging. We are looking for comment on these cases, good or bad. Suggestions always welcome! Updated with 'Final' version, February 2012

(CFP) Campus Bridging Early Adopter Call For Proposals

2011-11-30T20:56:42Z

For folks who are interested in the Early Adopter Program for Campus Bridging, we have additional information about the program and a sample proposal for people to look at.

The dates for sending Early Adopter proposals to campusbridging@xsede.org are Dec 1-9. Winning proposals will be notified by Monday, Dec 19.

Proposals will be evaluated with the following criteria:
Technical merit (corresponds to intellectual merit of NSF proposal):

Will the proposed project provide a good and thorough test of the capabilities that XSEDE GFFS offers?
Is the proposing team well qualified to offer critiques back to XSEDE about GFFS and the GFFS installers?
Does the clarity of the proposal instill confidence that pilot project can be well executed on the proposer's campus?

Broader impact:

Will the proposed project likely lead to broader adoption of GFFS on campus - and how many end users might be affected?
Does current/past use of XSEDE (and before it TeraGrid) suggest that the pilot will lead to immediate impact on current users of XSEDE?
And if not, is there a solid case that this pilot will lead to broader use of XSEDE on the proposing campus?
To what extent does the proposing team anticipate serving students from traditionally underserved groups (as defined by NSF)

Proposals are expected to deliver actionable user experiences back to the Architecture team within 6 months of initiating the Early Adopter Program.

For information about what should go into your Early Adopter proposal, please see the attached sample. You can also see a copy of the original CFP.

Year One priorities

2011-10-14T18:59:06Z

What do you think the Year One priorities of the XSEDE Campus Bridging effort should be? This is not a large effort in terms of personnel, so we're looking for what you think the essential activities should include. The effort is currently staffed by fractions of two people(Craig Stewart-IU and Jim Ferguson-NICS) along with some unfunded assistance from Therese Miller and Rich Knepper, both at IU. Please give us your guidance!

Campus Bridging @SC11

2011-10-14T18:53:47Z

The XSEDE Campus Bridging team will be available at an informal BOF at SC11. Please join us! This is scheduled for the WSCC, Room 3A/B, from 3-4pm on Tuesday (11/15). This is not an "official" SC11 BOF so it will not appear in the conference agenda.

We plan to give a brief update on where plan and activities stand, answer questions on the XSEDE Bridging effort, and promote the Campus Bridging RFP for those campuses who are interested in helping to be pathfinders.

What is campus bridging to you?

2011-10-11T22:09:38Z

Below is a draft explanation of campus bridging.

Please ad your comments about what campus bridging means to you!

What is Campus Bridging?

The challenge
Researchers, educators, and students have access to a wealth of digital resources, from the desktop to the campus to large national programs like XSEDE.

Currently it can be challenging to move among these various “cyberinfrastructures.” A chemist may become an expert at using her campus-level resources, but feel as though she’s starting from scratch when she’s ready to scale up to use larger resources through XSEDE. And her challenges will not be precisely those faced by a colleague at another institution, because each campus infrastructure is different—and researchers are moving to XSEDE from many, many diverse local systems.

How do we lower adoption barriers and build bridges so computational scientists and engineers are empowered to use any of the resources at their disposal?

The goal of XSEDE’s Campus Bridging initiative is to make it easier to leverage diverse computing resources and various cyberinfrastructures.

“The idea is to allow any researcher to use local cyberinfrastructure and XSEDE as though they were peripherals to their laptop.” -- Craig Stewart, Indiana University

Not just a project, a worldview
Campus Bridging is more than a single project—it’s a way of thinking about access to digital resources. XSEDE Campus Bridging co-leads Craig Stewart and Jim Ferguson will work to keep the entire XSEDE team thinking about this philosophy: Moving between local and national resources should not be an insurmountable challenge for researchers but instead should be a smooth path to increased productivity.

Campus Bridging efforts will include:
• Working with the XSEDE Architecture team to pilot use of GFFS, which enables communities to plug their file-systems into an XSEDE-wide global file-system that can be transparently accessed from XSEDE and non-XSEDE resources.
• Advising campus-level IT staff on configuring their local resources in ways that will improve the transition to XSEDE
• Helping Campus Champions to effectively communicate with their campus colleagues
• Serving as a conduit for feedback to XSEDE leadership