Kevin Noreen is senior manager at Dell (www.dell.com) for their Open Systems Product Marketing that focuses on HPC. Acknowledged as a visionary by his peers, he has been instrumental at Dell in identifying emerging markets and trends, providing strategic direction, developing leading partnerships as well as defining Dell’s future solution products and delivery strategies. We asked him how he would respond to the question that modeling and trading platforms have not always been viewed as part of the traditional enterprise applications.

KN: Modeling and trading platforms and applications hint at being HPC or compute-intensive versus data-intensive, so people have perceived the workload and supporting technologies to be outside of the more commercial mainstream, datacenter, disciplines, database and OLTP. What’s interesting is that without the HPC disciplines, the Street couldn’t perform all of the compute-intensive applications such as actuarials, derivatives, equities and municipals. Monte Carlo, portfolio management, bond speculation and net present value are all important to trading firms that depend heavily upon time, speed and performance which couldn’t be achieved without the high-performance systems and supporting software behind them.

WFS: Even with recent gains in the price/performance attributes of high-end workstations, the data-intensive nature of most modeling and trading apps require scalability beyond what is available in a single workstation. What solution does Dell offer?

KN: Companies are offloading their workgroup and departmental applications to cluster-based solutions to enable more available cycles on the desktop while the cluster does all of the heavy compute. This frees up the desktop to run other business-critical applications so traders can start plugging in data to their spreadsheets to deliver faster response times to their customers, by having the back-end cluster parse through the analytic workloads that otherwise take hours or days to perform on small systems.

WFS: What’s Dell bringing to the marketplace?

KN: A complete solution. Starting with our Precision Workstations, when users need more performance, they can seamlessly migrate workloads to high-performance, highly available clusters as our cluster solution leverages the best of commodity. Systems can be configured with Intel or AMD processor technologies running Windows or Linux or both. We also provide entry-level cluster solutions, from four to 16 nodes, to very large-scale systems up to 256 nodes or more. Through our partnerships with Microsoft and Platform Computing, we also offer solutions leveraging Windows Compute Cluster Servers and Excel 2007 as well as Platform’s Symphony to identify calculations within spreadsheets farmed out to the cluster, then returned to the financial spreadsheet, significantly reducing calculation time. We’ve worked with a number of customers who’ve experienced a 10-fold reduction in calculation time with a small 8-node cluster.

We are also seeing significant interest around Windows Compute Cluster Server technologies because CCS provides a natural migration path for pure Windows-based environments. For many of the applications that were previously being run on a desktop, CCS users aren’t even aware that they’re running on a cluster. It’s totally transparent.

WFS: With Dell and Microsoft working together in the HPC space, how are both firms combining lessons learned delivering high-end compute clusters with the ease-of-use requirements demanded by financial services managers?

KN: Windows CCS offers a simple path to high-performance clustering. For organizations with Windows expertise, the move to CCS is a more logical extension, especially for firms with limited in-house IT staffing or expertise. Microsoft is spending a lot of time working with Dell, to make sure we deliver tiered solutions capable of meeting every level of customer requirements – small, medium, on up to large-scale businesses, applications and requirements.

Time is money on Wall Street and why we aim to decrease the time between when a job is submitted to when it is complete.

WFS: And clusters can get you there….

KN: Easily and naturally.