AEROSPACE

Rogue Wave Software has teamed with other commercial companies and government agencies to tackle a challenging computing issue for the Air Force Space Command Project in a way that has turned the notion of high performance computing on its ear. As a part of a challenging risk reduction prototype effort, Rogue Wave is working with one of the Federally Funded Research and Development Centers (FFRDC), the MITRE Corporation. Rogue Wave and MITRE are demonstrating that functionality that has required a large mainframe environment due to the complexities of the processes can now be managed with an Intel Xeon processor-based server environment and the right software. The challenge with moving from the mainframe to a network of smaller, less expensive hardware is that the smaller servers get much of their computing power from "multi-core" chips, which the legacy software was not architected to leverage. The prime integration contractor for the Air Force's Integrated Space Command & Control program, Lockheed Martin, is working with the FFRDC and using Rogue Wave Hydra to help simplify the migration to the Intel multi-core servers. Lockheed Martin will migrate this prototype system to an operational capability for Air Force Space Command in the very near future. More specifically, Air Force Space Command (AFSPC) needed to move an existing space algorithmic application to a new architecture that could support a massive computational workload representing a large number of observations that included differential corrections. In fact, the specified growth requirement for transactions per second for this project exceeded the current world record for centralized database input/output. As a result, the new architecture needed to address millions of independent processes, concurrently, with maximum throughput and scalability, by leveraging newer multi-core hardware which can readily divide the problem across processors to achieve quicker results. Using Rogue Wave Hydra, the combined commercial and government team was able to migrate away from a large and expensive mainframe environment to the Intel-based hardware cluster running multi-core processors. As a result, the team has deployed a highly scalable solution that has greatly reduced the amount of work its developers need to perform, allowing them to concentrate more on enhancing the functionality. Considering the size, scale and speed of the solution, the most significant benefit may be the cost to build and maintain it. "The entire solution, including hardware and software, was a tenth the cost of running it on the old mainframe system for just one year, and the old system struggled to keep up with the load," said a senior U.S. government team member. "This project is a milestone in achieving price/performance advancements. But it also demonstrates the potential in migrating existing solutions to a multi-core architecture," said Tom Gaunt, senior vice president worldwide sales and marketing at Rogue Wave. "Any dynamic enterprise facing increased performance requirements can extend the value of their existing C++ investments if they use the right parallel computing solution."