Following the formation of a strategic partnership between Xenomorph and Microsoft last month, the vendors have now completed a project to integrate Xenomorph’s time series technology with Microsoft’s SQL Server 2008. However, the partnership was formed to apply high performance computing (HPC) to data management in financial markets and the vendors are also engaged in integrating Microsoft Windows HPC Server 2008 technology with Xenomorph’s TimeScape data and analytics management system.
Brian Sentance, CEO of Xenomorph, explains the background behind the partnership: “Xenomorph’s partnership with Microsoft has got stronger over the past few years, in particular with our joint work and cooperation going on with Excel Services (server-side spreadsheets) and most recently on implementing a time series database from within SQL Server.” He continues: “In relation to HPC specifically, then we have been involved with them over the past 12 months, starting with some joint work on a whitepaper. The partnership is non-exclusive, however Microsoft were chosen first due to our joint interest in applying HPC technology to financial markets, particularly in the field of data and analytics management. Our architecture is being designed to allow other HPC solutions to be used with TimeScape in addition to Windows HPC 2008, such as those from Platform and DataSynapse.” The vendors indicate that the partnership is a response to expanding data volumes, the pressure for real-time data management and increasing regulatory demands. Under the HPC agreement, Xenomorph will apply Microsoft’s HPC technology to the intraday calculation of large correlation matrices, a key calculation for correlation-based trading and risk management. Thus Windows HPC Server 2008 will become the first HPC platform to be fully integrated within TimeScape. Xenomorph reckons this will deliver scalability in reference and market data management, allowing TimeScape to analyse high volume market data and accurately value derivative instrument portfolios. Sentance’s colleague Chris Budgen, chief technical architect at Xenomorph, explains: “Faced with an overload of data, technology specialists and business users need greater scalability from data management systems. Data and its analysis are key pressure points and are growing as data volumes challenge traditional data analysis and instrument pricing. We need to move away from inflexible, monolithic data management architectures to a more dynamic and service-based approach.” Michael Newberry, HPC product manager at Microsoft UK, adds: “In today’s business environment, people are looking for computing systems that are scalable and responsive to their business needs. With partners like Xenomorph onboard, Microsoft Windows HPC Server 2008 is providing a responsive and cost effective alternative to UNIX and mainframes. Windows HPC Server 2008 enables businesses to meet the computing challenges, and the data volumes, of the 21st century.” Sentance believes that data has run ahead of the market’s ability to analyse it: “Even with high volume storage of real-time market data, when it comes to data analysis, many institutions find themselves asking the question, ‘We’ve got the data, now what?’ Parallelisation using HPC is the only way to keep on top of this and helps competitive advantage in bringing new trading and risk management ideas to market.” He reckons that the architecture of some enterprise data management (EDM) solutions can be somewhat monolithic and inflexible. “So Xenomorph’s initiative with Microsoft to introduce HPC technology to data management is about offering a more agile, fault tolerant and cost effective architecture that will scale with the performance needs of our clients,” he continues. “On a related note, the usage of HPC technology also reflects the risk management need to manage data, calculations and pricing together within a consistent and transparent framework. Calculated or derived data is very valuable to our clients and Xenomorph believes this should be actively managed in just the same way that ‘raw’ market and terms and conditions data is for instruments,” he adds. Sentance doesn’t think there will be any significant challenges in integrating the two vendors’ solutions: “The existing architecture of our data and analytics management system, TimeScape, is well suited to componentisation which is a good start when re-architecting for use within a HPC architecture.” With any HPC solution, a key issue to be addressed is to look at how data is marshalled within the cluster of compute nodes and in this respect Xenomorph is currently looking at distributed cacheing technologies such as those available from GemStone and GigaSpaces, says Sentance. “Ultimately, the movement to service oriented architectures (SOA) involving HPC and grid technologies is a big shift for many institutions, so ease of deployment and management is another key focus on this initiative,” he explains. Sentance concludes: “The Xenomorph approach is all about the management of both data and analytics, providing agile but controlled access to any piece of market data, reference data, statistical calculation or pricing model. Many systems in the market are unfortunately siloed by type of data, for example, market data or reference data or derived data, and do not have the flexibility to support more complex products and the pricing models that go with them. Put another way, there needs be greater convergence in the way that data is managed today. Put in the context of current market conditions, risk managers need tools that provide them with both the power and flexibility to look at any kind of data, in any way that they want to. Transparency has never been so valuable and important, and this is Xenomorph’s aim in delivering this capability to all of our clients.”