Laboratory of Neuro Imaging, UCLA

Brain Research Lab Makes Smart Choice with Sun Servers

Begun in 1983 at the Washington University School of Medicine at St. Louis, the Laboratory of Neuro Imaging (LONI) moved to the University of California at Los Angeles (UCLA) in 1987. Today it is one of the country's foremost neurological research centers. LONI is a leader in the development of advanced computational algorithms and scientific approaches for the comprehensive and quantitative mapping of brain structure and function.

Business Issues

Ensure high level of productivity for researchers now and in future
Shorten execution times of critical applications
Maximize utilization of computing resources

Solution

LONI deployed a 306-node cluster of Sun Fire V20z servers running Sun N1 Grid Engine software, which manages the lab's servers as a computing service. Each Sun Fire V20z server has dual 64-bit 2.4 gigahertz AMD Opteron processors with four gigabytes of memory.

Business Results

Increased productivity for researchers due to increased computing power and faster application execution
Ability to scale to accommodate growth
Increased utilization of computing resources

Products / Solutions

Success at a Glance

The human brain is the ultimate computer: By one estimate, its computational power is equivalent to 100 million MIPS. Understanding this powerful and complex organ holds great promise for treating and even curing a wide range of brain diseases such as Alzheimer's and schizophrenia.

The Laboratory of Neuro Imaging (LONI) at UCLA is working to make that promise a reality. LONI's unique contribution to neurological research is the creation of digital atlases drawn from thousands of scans of individual human brains. The associated collection of data is enormous, o ver 30 terabytes - and so are the computational challenges of analyzing that data.

Recently, Dr. Arthur Toga, the director of LONI, realized that the lab needed a new computing system. The number of researchers and the amount of data were growing, straining current resources. Those researchers are usually neurobiologists, not computer scientists, so they need sophisticated analysis tools that are also easy to use.

“ Our Sun solution helps researchers work faster--and better. Data analysis that would have taken days now takes hours, making it possible to pose more complex questions, and get answers. Now, even a neurobiologist with little computer training has full access to the lab's vast computing power, thanks to the Sun grid solution. ”

— Dr. Arthur Toga, Director, Laboratory of Neuro Imaging, UCLA

LONI invited a number of vendors to compete for the business. The key part of the competition was a series of 12 tests involving mathematical algorithms such as solving a set of simultaneous partial differential equations. The vendors were evaluated on both compute time and "wall-clock" time.

Sun had the highest performance using Sun Fire V20z servers with dual AMD Opteron processors and the Solaris 10 operating system. In addition to the good relationship between Sun and LONI, dating back 10 years, Sun also received high marks for technical support and commercial stability. In the summer of 2005, LONI awarded the contract to Sun.

The Sun deployment includes 306 Sun Fire V20z servers running Solaris 10. Dr. Toga worked closely with the Sun Solution Center in Hillsboro, OR, to evaluate operating systems. While he was originally considering Linux, the benchmark tests showed that Solaris 10 would provide the best price/performance for LONI.

The servers are organized as a single massive cluster managed by Sun Control Station software. Sun N1 Grid Engine software manages the compute workload not only of the Sun Fire V20z servers but also legacy servers from other vendors as a single computing resource.

Researchers are benefiting from the new system. Because they get answers to their analyses faster, they can pose more - and more complex - questions in their available research time. For example, LONI researchers are currently investigating the effect on the brain caused by schizophrenia and AIDS. This research speeds the development of new treatments, relieving suffering and saving lives.

As a result of the Sun upgrade, LONI can support the computing needs of its current research community, and can easily scale to accommodate growth. LONI is evaluating other Sun servers, including next-generation x64 servers, targeting higher levels of energy efficiency and a smaller datacenter footprint. No matter what direction LONI's research takes, Sun will be part of the picturefor years to come.