Customer Snapshot: Education

University of California, San Diego

Cutting-Edge Research Organizations Advance Scientific Goals with Sun Fire X4500 Data Server Technology

University of California, San Diego (UCSD) is a powerful magnet for those seeking a fresh, next-generation approach to education and research. With a campus enrollment of 26,876, UCSD has rapidly achieved the status as one of the top institutions in the nation for higher education and research since its founding four decades ago.

Customer Challenges

  • Provide fast, 24x7 access to data for worldwide community of researchers
  • Manage large, rapidly expanding volume of research data cost-effectively
  • Minimize facilities requirements for power and cooling
  • Accommodate range of complex data types within one unified storage system

Solution

Modular, “plug-and-play” architecture for high-volume storage implementation providing 24TB of storage per four rack units, combined with powerful four-way Opteron processors and a high data integrity file system.

Business Results

  • Low power and cooling requirements due to ultra-dense storage technology
  • Up to 590MB per second data transfer rates clocked between X4500 servers and clients
  • High availability due to snapshot and replication technology
  • Ability to scale to petabytes with same infrastructure

Story Details

UCSD ranks among the top 10 universities in the United States for excellence in scientific research―far above many universities which are much older. One reason for this has to do with the university’s ability to develop and nurture organizations that foster interdisciplinary breakthroughs which enhances UCSD’s ability to compete for grant money.

Recently, the university turned to Sun to provide a storage infrastructure for research projects in two fields: environmental studies and marine biology. While the projects had different goals and requirements, the Sun Fire X4500 data server proved to be the cost-effective, scalable answer for both.


" We selected an emerging Sun product as a cost-effective way to deal with large amounts of data. We built our solution on the Sun Fire X4500 server and are very pleased with the result. "
— Mark H. Ellisman, Ph.D., Director of the Coordinating Center for the Biomedical Informatics Research Network (BIRN), UCSD

The Sun Fire X4500 servers are dual-socket, four-way AMD Opteron 285 machines, with 16GB of memory. The ultra-dense storage technology spares power and cooling resources and the 4U size conserves floor space. Researchers can retrieve data quickly regardless of the number of data types involved. In addition, ZFS, the 128-bit file system of the Solaris 10 Operating System, provides a mechanism to store vast quantities of information within a single file system and naming space for each X4500 server.

The UCSD project that first benefited from the Sun Fire X4500 server is the Hurricane Response Portal, an interdisciplinary study headed up by Mark H. Ellisman, Ph.D., director of the Coordinating Center for the Biomedical Informatics Research Network (BIRN).

Using satellite imagery and mapping, the Hurricane Response Portal studies the environmental after-effects of Hurricanes Katrina and Rita to discover the extent of toxic pollution and information that might help future evacuation and relief efforts. The project is funded by the National Institute of Environmental Health.

Another data-rich project at UCSD is the Cyberinfrastructure for Advanced Marine Microbial Ecological Research and Analysis (CAMERA) project. CAMERA provides the largest available set of gene data to researchers worldwide. These scientists study the genetic makeup of microbial life found in samples of seawater, or other environmental niches, taken from around the world. In addition to providing the most comprehensive set of genomes of a single microbial species, the CAMERA project provides “metagenomic” DNA reads representing a vast number of organisms which live together, revealing their relationships to each other and to the environment. Its principal researcher is Dr. Larry Smarr, director of the California Institute for Telecommunications and Information Technology (Calit2). CAMERA was funded by the Gordon and Betty Moore Foundation.

A key figure in developing the cyberinfrastructure of both projects is Philip Papadopoulos, program director, Grid and Cluster Computing, San Diego Supercomputer Center, UCSD. An expert in building high-performance storage infrastructures for distributed front ends, Papadopoulos, along with Ellisman, had followed Sun’s development of the technology underlying the X4500 server even before it was released and were convinced it represented a breakthrough.

The infrastructure needs of both projects were similar: store vast amounts of data in a number of data types, and provide the data quickly in response to queries from researchers. To keep power and cooling requirements down and conserve fast-disappearing floor space on campus, UCSD needed an easy-to-manage, modular, shared storage solution rather than an expensive and high-maintenance storage area network (SAN) solution. Together these projects use more than 400TB of raw storage in Sun Fire X4500 servers. The data servers are connected to the network with 10Gb Ethernet. Data transfer rates between the Sun Fire X4500 and clients have been clocked at as high as 590MB per second.

For both projects, Sun provided professional services to help with installation and deployment.

Using Sun data server technology, research projects such as these can reach their goals rapidly because the infrastructure can be assembled in a matter of weeks. To shorten the time-to-research even further, Papadopoulos and Smarr chose Sun Customer Ready Systems program to rack, stack and cable the servers and make them production-ready. Now UCSD scientists and other researchers on the cutting edge can perform research with an infrastructure that keeps pace with their needs.

  
 
Interested in Sun's Open Storage?
Download this paper today to learn about the tools, trends and key features of Sun's Open Storage solutions.