Andrew Meehan
It's common knowledge that BLAST's Web interface puts the power of the National Center for Biotechnology Information (NCBI) on users' desktops. But what exactly does that mean?
To the user, the process appears seamless: submit a query, wait a few moments, and an answer magically appears. On the backend, however, the process is more complicated. The NCBI receives nearly 200,000 BLAST queries per day, and it devotes some 130 dual-CPU computers to the task of working through those requests.
As NCBI's David Lipman explains, those requests go to the BLAST frontend, which insert them into a structured query language (SQL) database. The SPLITD server daemons then pick up the requests and split them across the BLAST backends ("BLAST Servers"). Separate pieces of the query can thus be processed on up to 10 servers simultaneously.
Most of the sequence data (including GenBank and other sequence databases) is held in ...
