Case Studies

Georgia Institute of Technology

Universities have a long history of using high-performance computing (HPC) for research and
analysis in engineering, physics and mathematics. In recent years, however, high-performance
computing clusters have also revolutionized traditionally descriptive sciences such as biology,
with the ability to dedicate the power of literally thousands of processing cores to simulations
that seek to unlock some of the deepest secrets of our bodies.

University of Virginia

Researchers have used crash test dummies in vehicle rollover tests for years to help them determine why these events have a higher fatality rate than any other type of crash. Now they are turning to high-performance computers, which researchers say are more practical and cost-effective when it comes to pinpointing elements in future automobile designs that could reduce injuries and deaths.

Digg

Digg is well known as one of the leading innovators in the evolution of the Web. Digg enables people to discover and share content from anywhere on the Internet by submitting links and stories and voting on them. The success of Digg has helped spur the growth of social networking worldwide.

Plaxo

Plaxo started as a different kind of address book, leveraging the power of the Internet to help customers connect and stay connected with friends and family, but it has quickly grown into a leading online address destination for millions of people. In 2006, Plaxo took the notion of staying “connected” a step further and partnered with America Online (AOL) to integrate Plaxo’s services with AOL Instant Messenger, essentially doubling its user base almost instantaneously. At that point, Plaxo realized it would need to overhaul its IT infrastructure to support a major influx of users and their data on a tight timeframe—it had just four months to complete the upgrade.

First Victoria National Bank

Texas-based First Victoria National Bank (First Victoria) had been growing steadily since it opened its doors in 1867. But, since they never do anything small in Texas, now it wanted to really grow – more than double in ten years, to a $2 billion dollar company sprawled across Southwest Texas. That was a real wake up call to the information technology (IT) staff who already struggled with an IT system that was not getting the job done for the bank’s current needs. But the bank planned ahead and invested in an IT overhaul based on Penguin Computing Relion servers. With Penguin Computing’s help, the IT group streamlined their cross-platform infrastructure elements into an efficient, scalable system that virtually eliminated email and customer relationship management (CRM) database downtime. With no set-up or implementation issues and no problems since deployment, the First Victoria team also has the scalable IT tools it needs to meet the future banking needs of the state.

The Planet

A leading server-hosting company, The Planet has over 200,000 square feet of dedicated and co-located datacenters. When it came time to expand its array of products to help The Planet meet the growing demand for dedicated enterprise and gaming servers, The Planet chose to add a new line of servers from Penguin Computing for the fastest, most powerful server capability The Planet has ever offered.

National Weather Service

To issue up-to-the-minute weather updates and warnings to the public for a region covering everything from Albuquerque, N.M. to San Juan, Puerto Rico, the National Weather Service (NWS) Southern Region Dissemination Enhancement Team needed reliable, powerful computing technology. Behind the scenes, enormous quantities of data from satellites, marine buoys, and many other sources had be analyzed and dynamically updated to support heavily used applications such as the Graphical Point Forecast (http://www.nws.noaa.gov/) website, which averages 15 million hits a day. Stable Penguin Computing servers and Scyld Beowulf™ clustering software got the job done by achieving the sheer computing power needed and helping NWS stay within their budget. Using this robust combination of products, the NWS Southern Region team improved the quality of the data it presents to the public, in line with the NWS mission of “producing and delivering information you can trust when you need it.”

Sandia National Lab

Sandia National Laboratory’s Combustion Research Facility (CRF) had a mission of finding high efficiency, low emission solutions to complex combustion problems for power plants, internal combustion engines, industrial furnaces and other applications. But their simulations for this research were so complicated and had so many variables that existing computers at CRF could not run the jobs. And even though CRF was doing science for the Department of Energy with real world pocket-book impact, they were often “last in line” among other government facilities, academic researchers, and other applicants to get access to the supercomputers they needed. Penguin Computing and Scyld Beowulf™ clustering software solved the problem by cost-effectively building a Linux cluster that could perform most routine calculations, leaving only the more complex simulations for supercomputer requests. Within one day of the equipment’s arrival, the cluster was up and running and has gone to perform so well that CRF eliminated the need to budget $150,000 per year for support staff. Instead, CRF now regularly runs calculations on their cluster that in the past would have required dedicated, massively parallel computing power using 1,000 processors!