From portfolio managers to risk managers, an incomplete picture of data can produce an inadequate result. As firms extend their lines of business, they add more silos for data. Without an enterprise-integration solution, those silos remain out of reach for many of the functions that would benefit from their data contributions. Moving beyond data warehousing, the industry is now at the next evolution of data management, explains Gene Kim, a senior analyst at Financial Insights. "Now people are becoming more aware of broader enterprise-wide implications of data," he says.
The concern for data integration is beginning to transcend the data world and become of interest to upper management, Kim adds. "There are enterprise risk-management issues, where C-level officers can't reconcile risk reports and financial data that is coming from different stores and systems."
Chadwick Nestman, chief architect for JPMorgan Investor Services and lead architect for the Treasury and Securities Services, says that customers are also driving data integration. "Companies are finding a tough time bringing all the data together to allow the customer to get comprehensive access," he says. "That's where the integration comes in."
Yet, despite the sparks of interest, Wall Street has yet to see complete enterprise integration of data in widespread deployment due to what Nestman sees as a lack of funding. "Many businesses aren't willing to spend money on [middleware or infrastructure], given what the economy's been, to work on something that has no direct value to the customer," he says. "But there are some companies that have made that jump, and they're already beginning to reap the rewards from it."
Guardian Group of Funds, a subsidiary of Bank of Montreal, is beginning to see benefits from its recent enterprise-application integration (EAI) of its databases. After investing in a customer-relationship-management system from Siebel a few years ago, the firm began examining ways to tie information together to avoid data redundancies, says Julie Hodge, assistant vice president of Information Services at GGOF.
Hodge describes the firm's ideal solution as one that could provide a data warehouse as well as an enterprise-wide reporting solution that could connect with both the legacy applications as well as the new data warehouse. In addition, the solution needs to have an extract, transform and load component in order to move data from one system into another. Lastly, she says, GGOF wants the ability to publish its data into online portals.
GGOF eventually decided to implement WebFocus, a hub-and-spoke data-warehouse model from New York City-based Information Builders. Information Builders migrated information from three legacy systems (CRM data from a Microsoft SQL Server, fund-accounting data from Sybase and record-keeping data from DB2) to create a new, central warehouse residing on a Microsoft Windows NT server.
While the central warehouse holds data pertinent to all lines of business, Hodge specifies that GGOF did not move all data into the warehouse. "You don't just want to copy data for the sake of copying it," she says. "I wouldn't take a lot of transaction-based information off my record-keeping system and store it again in my data warehouse." Kim agrees that some data should remain in outlying databases. "There are pieces of data that are only relevant to a couple of applications or processes. "The key is that you have to know where that data is, what format it is in, and how to access it."
In addition to warehouse and legacy-system connectivity, the data can be accessed using a Web portal. Users can log onto the system and monitor the data via a key-performance-indicators dashboard. "It allows users, within any line of business, a one-stop shop portal to access the data for their reporting requirements or informational needs," says Sean Allan, the enterprise-business analyst at Guardian Group. "A .NET platform is going to allow us more flexibility to get into Web services."
Allan adds that the enterprise integration of GGOF's data is rolling out in phases. While GGOF is relatively small in size, managing only $2.8 billion in assets, Hodge concedes that for a large firm, an EAI project can be costly and time-consuming.
While Guardian opted to implement an EAI solution, a newer concept is making its way into discussions on data management: Enterprise Information Integration (EII). EAI examines how one can link applications to speak to each other, while EII focuses on making the data available across diverse environments and applications.
"People are starting to create a virtual warehouse where you can just publish the data, and whoever wants the data can just ask for it," says Scott Dillman, a partner at PricewaterhouseCoopers, describing EII. "There is no big container of the data; it's just out there on the intranet."
JPMorgan's Nestman adds, "Instead of running an application, you're accessing the data as of the latest time it's been updated. So the user gets a direct link into the data and the user-reporting engine. You can slice and dice it, reshape it, tag it or build tables. Whatever you want to do with it, you can." In order to implement an EII solution to make data universally available, though, a firm needs to have common definitions of data elements. Nestman says that tools from EII vendors, such as New York City-based MetaMatrix, use a layer to translate data from legacy applications into one common language. Tools such as these use metadata databases, containing data that describes data, in order to define data elements in the context of their usage.
While standards such as XML-based eXtensible Business Reporting Language are ideal, Nestman contends that legacy systems make it difficult to convert existing databases into one common definition language without a translation layer between the data and its recipient.
Some firms, such as Merrill Lynch and Bank of America, are beginning to employ EII models to address data integration. For others firms on Wall Street, such as Guardian, Web services seems a bit farther off. "I don't see [Web services] coming to the forefront for at least another 18 months," Guardian's Allan says.
Despite the confidence PricewaterhouseCoopers' Dillman has in the future of EII and Web services for enterprise integration of data, he points out that EAI isn't going to disappear. "Big data warehouses aren't broken," he says. "For legacy systems, people still need EAI, but for the next level, service-oriented integration is more tactical."
"We're just at the beginning of looking at EII," Nestman says. "Over the next two to three years, we're probably going to see a greater emphasis on EII. Then in five years, you're going to see the cap come off, and everybody will have major programs going down that path."
Julie Hodge, assistant vice president of Information Services at Guardian Group of Funds, offers this advice for companies taking on enterprise integration of data: