Frontier Science Undergoes an IT Platform Upgrade
How to Increase the Reliability of Your IT Infrastructure Using Predictive Analytics REGISTER >
A leading research foundation that focuses on cancer and AIDS undergoes a major IT platform upgrade to boost business performance.
Few organizations face greater data management and processing requirements than top-tier research foundations. The need to sort through huge volumes of data—and handle the task relatively quickly—is at the center of achieving real-world results. At Frontier Science and Technology Research Foundation, a not-for-profit organization that manages clinical trials for HIV, breast cancer, tuberculosis and other key areas, the need for a more advanced technology framework has led to significant upgrades and changes over the last couple of years.
"We are highly dependent on advanced servers and storage to process data," said IT director Peter Meszynski.
The company, which focuses heavily on software and general computing techniques specifically tailored to the needs of data collection, categorization and analysis, had been relying on a SPARC Solaris-based and x86 computing platforms for about 15 years. As the most recent generation of servers approach the end of their life, the foundation began searching for new systems that could accommodate rapidly growing workloads and data requirements. However, "We wanted to keep both environments in place," Meszynski said.
In fact, hosting the most current group of existing SPARC Enterprise M5000 servers on an external site would have been cost-prohibitive, he said. Based on processing power, energy consumption and cooling requirements, Frontier Science, which operates offices in New York, Massachusetts and Wisconsin, opted to go in a different direction. In January 2015, after about 6 months of research and testing, it installed Fujitsu M10-4 and M10-1 servers at multiple sites along with Fujitsu ETERNUS DX200 storage solutions. The smaller server footprint not only made it easier to install the systems, it led to significant improvements in energy consumption.
According to Meszynski's calculations, a single M5000 pulls about 9.26 Amps at a sustained rate, while the new Fujitsu M10-4 uses only about 4.2 Amps. The more than 60 percent reduction has contributed to significant cost savings across the co-located environment. But the platform has also helped Frontier Research boost throughput by 50-70 percent, and ensure optimal availability of the servers along with mission critical services, tools and data to a staff of about 250 and more than 6,000 researchers located around the world—all on a 24/7 basis.
Meszynski added that the gain in computing power also allows the company to "run considerably more routines simultaneously."
The upgrade did not require any recompiling of code. As a result, migrating data and critical applications such as SAP Crystal Reports and Actian RDBMS was simple and straightforward.
So far, the platform has provided 100 percent reliability, Meszynski said. What's more, the combination of Solaris zones and the SAN server's snapshot feature allows Frontier Science to verify replications and restart all replicated machines weekly in a sandbox.
"This is incredibly useful in verifying that the disaster plan will work," he said.
Because it is using only about 50 percent of the Fujitsu M10 server cores, there's plenty of room for growth and a lot of flexibility in terms of how the organization approaches computing requirements. The company's data volume is growing by about 15 percent annually.
"This ability to grow…means the platform will have a significantly extended lifecycle and return greater value for our money," he said.