Taking Control of the Cluster

When Gernot Stocker noticed that the biologists in his lab were not using the department's brand-new, 24-node computer cluster, he started taking notes. Running sequences on a desktop computer could take days, yet these scientists preferred that to the cluster for a simple reason. "Biologists don't like command-line software," says Stocker, a graduate student at the Institute for Genomics and Bioinformatics at Graz University of Technology in Austria. "They're used to using Web browsers."

So Stocker programmed a simple solution: a Web browser-based interface that will run most sequence-crunching algorithms on the master PC that controls a Linux cluster. Called ClusterControl http://genome.tugraz.at/Software/ClusterControl, this queuing program uses any Microsoft, Netscape, or Mozilla-based browser.¹ It requires installation of PHP (a server-side scripting language), Apache (a Web server), and OpenPBS (a queuing system).

In just three steps, a user can have the cluster run a sequence using NCBI BLAST, FASTA, or any of ...