Fusion-io at UC San Diego for MySQL Database Queries

Fusion-io, Inc announced that its technology has been utilized to realize improvements in MySQL database queries for bioinformatics research.

The Protein Data Bank (PDB) research is being conducted at the University of
California, San Diego, in collaboration with the San Diego Supercomputer Center
(SDSC).

Researchers at SDSC noted that replacing HDDs with
Fusion-io technology in their database infrastructure reduced query times from
30 minutes to three minutes

Fusion’s ioMemory technology is being utilized by the SDSC,
an organized research unit of UC San Diego, in its data-intensive computing
initiatives. As part of these efforts, researchers are using the technology to research on the Protein Data Bank, a
repository for three-dimensional structures of large molecules and nucleic
acids. The research carried out for the PDB serves as a foundation for
accelerating the development of life-saving drugs, synthetic proteins used in
medicine and other treatments for illnesses.

"As SDSC strives to serve as a center for research
excellence, we focus heavily on evaluating how flash-based technologies can
improve scientific discovery," said Allan Snavely, Associate Director of
SDSC and co-Principal Investigator of the Center’s upcoming data-intensive HPC system, Gordon, which is currently in
development. "Through our research, we are finding that storage memory
products like those offered by Fusion-io are considerably faster than hard
drive disks for many large memory and data-intensive problems, and that
flash-based storage memory has the potential to make supercomputing-level
results accessible outside the traditional user base."

Using Fusion’s server-deployed technologies, researchers
were able to reduce the time required to conduct complex MySQL database queries
that determine relationships between proteins and the effects proteins can have
on each other. Using traditional hard drive disks, a query analyzing more than
200 million protein structures took 30 minutes.

"With Fusion’s shared data decentralization
technologies – which place the storage memory medium right next to the
processor to significantly reduce latency and increase processing speed – the
same query took just three minutes," said Robert Sinkovits, SDSC
computational scientist and Gordon Applications Lead.

Spencer Bliven, a graduate student in the Bioinformatics and
Systems Biology Department at UC San Diego working with Allen Snavely on the
research, said: "The PBD is helping us gain a more theoretical and
practical understanding of the foundations of biological science. With the
reduced query times I’ve seen in conducting tests with Fusion’s ioMemory
technology, I’m able to stay more focused on my research. This allows the
scientific discovery process to be more spontaneous and greatly helps me in
eliminating potential errors."

The computer used to realize the performance gains includes
two quad-core Intel Xeon E5530 2.40 GHz processors and 48 GB of DDR3-1066
memory. It includes four 320 GB ioDrives mounted and configured as a single 1.2
TB RAID 0 device running an XFS file system.

"Storing crucial data in the server where it is
processed can greatly accelerate the analysis of large data sets to enable new
scientific discoveries," said Neil Carson, Chief Technology Officer of
Fusion-io. "It’s inspiring to see how SDSC is leveraging the performance
gains and reduced latency made possible by Fusion-io technology to advance
complex biological research that serves as the foundation for many improvements
in science and medicine."