"Our clients have commented that they have experienced a price performance increase of five to ten times for the internal applications they have already ported to Linux," said Michael Zerbs, Chief Operating Officer at Algorithmics. "Through Linux and distributing support we can deliver the proven enterprise-wide risk management capabilities in the industry at a lower cost, while enhancing performance and scalability - a truly powerful combination."
Over the past year, Algorithmics' Linux and distributed computing initiative has made considerable progress including the initiation of two separate client pilot projects. In 2002, Algorithmics formed a key partnership with DataSynapse to incorporate their GridServer technology as part of the distributed computing version of Algo Suite. Algorithmics has also recently completed performance benchmarking with both IBM and HP that has confirmed significant enhancements in both speed and scalability. These tests also confirmed that additional computing resources can be added to the solution without extensive implementation or configuration work required, thereby reducing the total cost of ownership. Additional benchmarking exercises are being planned with Intel and Egenera and will ensure that a Linux version of Algo Suite will be supported across all major client hardware platforms.
With the release of Algo Suite 4.4 on January 15, 2004, it will be possible to implement and run Algorithmicsâ Mark-to-Future (MtF) architecture entirely on Linux and deliver advanced enterprise risk management across the banking (wholesale and retail) and trading books. Using a Linux grid is perfectly consistent with Mark-to-Future methodology. Generating the MtF cube decomposes naturally along the instrument dimension and can be easily split among multiple machines. Using GridServer from DataSynapse, the generation of the MtF cube can be split across multiple machines with each machine simulating a small part of the cube.