And no industry has been quicker to adopt each successive, productivity-enhancing wave of information technology, from large-scale mainframe computing in the 1950s and 1960s to distributed minicomputers and client-server networking in the 1970s and 1980s to the explosion of Internet applications and Web services over the past decade.
But now the financial industry -- and especially the top Wall Street banks, which have consistently pushed the limits of technology in their brokerage and trading operations -- is paying a price for all that progress.
The price is, simply, a back-office mess: a jumble of systems of varying ages that IT departments spend inordinate amounts of time and effort synchronizing. Demands on the tech infrastructures, meanwhile, have never been greater: They must be ready to capture the spiraling transaction and quote volumes generated by program and options trading while at the same time meeting new regulatory requirements to detect market timing and other illegal practices.
Now technologists are turning to yet another new set of solutions for their mounting systems-integration problems. These include IP (Internet protocol), XML (extensible markup language) and the Linux operating system -- all open standards that promise greater operating efficiency for a low price. Another option: grid, or utility, computing, which heightens productivity by linking processors and parceling out capacity as needed.
Are these the answers to the financial industry’s prayers, or will they just add another layer of challenges? If history is any guide, they are likely to be both.
“What this industry has done technologically over the past ten to 15 years is holding it back from where it needs to go,” observes Kenny McBride, who runs the capital markets business for Microsoft Corp.'s financial services group. “It’s almost as if it needs a clean slate.”
A clean slate is unrealistic, though McBride plans to organize a panel of securities industry executives to discuss the concept as an exercise. The question, he says, is, “How do we take current or near-current technology and make it last a generation?”
McBride is in sync with Nigel Woodward, worldwide manager of capital markets and securities at Sun Microsystems -- the Microsoft archrival-turned-partner that supplies high-powered workstations to many of Wall Street’s most sophisticated trading desks. “There was a cost of being early into client-server computing,” says Woodward, referring to the architecture that allows traders to execute database and analytic programs on their desktop, or client, machines. “They got powerful clients, but they are all islands -- thousands of PCs that have to be maintained and refreshed individually.”
Neither Microsoft nor Sun is pushing a single vision of the future financial infrastructure. They say they will respond to signals from the marketplace. Explains Woodward, “We’re a solutions facilitator, not a solutions provider.”
Confronted with the industry’s preference for easy-to-integrate, open-source standards like Linux in place of Microsoft’s Windows and Sun’s Unix, both vendors are in a position to facilitate one increasingly popular approach: grid computing. Until now it’s been mostly an entrepreneurial movement led by companies like DataSynapse, Egenera, Enigmatec Corp. and Platform Computing, which have developed grid programs for the likes of Credit Suisse First Boston, Goldman, Sachs & Co., J.P. Morgan Chase & Co. and Wachovia Corp.
The third major platform giant, IBM Corp., never inextricably linked with either Windows or Unix, has been quicker to champion both Linux and grid computing. IBM did pioneering work last year with Charles Schwab Corp. to assemble a grid to handle computation-intense risk and market simulations for investment advisers.
Typically, a “compute grid” shortens bulk-program running times from hours to minutes. But Geoff Penney, Schwab’s chief information officer, says the real payoff will come as the grid concept becomes more pervasive. “The way you establish an advantage is by continuing to build on an innovation step by step,” says Penney. “We are in an early stage of that process.”
Isn’t everybody? Extending the grid, or utility, concept enterprisewide will be the next big challenge, but it doesn’t require cleaning the slate, says Duncan Johnston-Watt, chief technology officer of London-based Enigmatec. “You don’t have to wait for a full-scale utility-on-demand,” he asserts. “The whole point of an enterprise grid is that you can do efficient computing now, with available resources.”
So financial firms must act fast -- before the next generation of solutions arrives with a whole new set of confusions.
