Searching biological sequence databases is one of the most routine tasks
in computational biology. This task is significantly hampered by the
exponential growth in sequence database sizes. Existing parallel sequence
search tools such as mpiBLAST have been focusing mostly on parallelizing
the sequence alignment computation. However, inefficient handling of input
and output data can easily create performance bottlenecks even on
supercomputers. In this talk I will present a set of techniques for
efficient and flexible data handling in parallel sequence search
applications, as well as some experiences and lessons learned from the
practice of large scale parallel sequence alignment.