Modern barn raising

This week at Fermilab we are building a new version of something called the open science grid integration testbed, which will eventually lead to the next release of the OSG production grid.  We are essentially try to work out all the bugs with software, installation methods, and validation tests.  We have a number of people from various VO's (virtual organizations) pitching in with the work - each expert and bringing something unique to the table.  Here is a snapshot of the current status:

Yes, this is how our computer grid begins its life - as a table on a whiteboard. As each site on the grid is installed, configured and validated, it gets a blue dot in the table.  You can see here we have a long way to go (today is the second day of the workshop, we have our work cut out.)

Over the years, we've done an exercise like this several times.  What I find interesting about the process is the community spirit that develops among the people working on the project.  There are about 30 people in a single room in the Feynman Computing Center at Fermilab, several others joining by Polycom, still others coming in through a chat room, each hunched over a laptop as they nurse their grid site or grid service into a functioning entity that science users "out there" will see compute and data resources appear on a map (or in something called an information service) and become available to submit their analysis jobs. "Hey, anybody know how to configure Monalisa?", or "Are srm clients required to be visible from compute nodes?", or "Whose turning up red in GridCat?", or  "Hey, can someone sign my host certificate?".  Something semi-magical happens when people, normally known only as a voice on a teleconference or salutation in an email materialize in person and are available for the quick consultations when the questions come.

My colleague Mike Wilde from Argonne Laboratory has often called this process "grid raising".  Seems an appropriate thing to be doing on the Fermilab prairie, next to the Buffalo herd.