By Tod Newcombe Contributing Editor In Massachusetts, the state's computers are tuned to process transactions, generate paychecks, perform general ledger and budget functions as well as run a host of other operations. The computers collect and store vast amounts of data, but are not designed to provide workers easy access to information. To remedy the situation, the commonwealth has built a data warehouse - a central, integrated database of information designed to support decision-making and analysis. Accurate, up-to-date information is fed regularly to the warehouse from the state's accounting, payroll cost reporting and budget systems. Analysts and decision-makers can use the warehouse to investigate cross-organizational issues, compare program finances with service outcomes and perform a host of other investigations. In the future, other state computer systems will feed their data into the warehouse, substantially increasing its size and scope. For the first time, Massachusetts' leaders have instant access to the kind of strategic information necessary to drive today's state government. The state did not purchase a single product called a data warehouse, because it doesn't exist. "A data warehouse is an architecture that must be built out of several key components," explained Kevin H. Strange, a research director at GartnerGroup, the market research firm. Data warehouse components and functions include: * Stores of operational data from legacy systems * Data conversion and extraction methods * A database management system * Warehouse administration * Business intelligence tools that help users find and analyze data contained in the warehouse. According to Strange, a completed data warehouse can benefit an organization by improving decision-making and enhancing customer service. They also can improve the quality of data for an organization, an often overlooked benefit. "A warehouse not only integrates data from multiple systems, but it cleanses the data," he said. Clean data ensures that an agency's multiple information systems complement rather than compete or conflict with each other. But in order for a data warehouse to succeed, it requires an enterprise strategy. According to Strange, warehouses are often mistaken for decision-support systems, "which help a line of business solve tactical issues. But data warehouses support strategic decision-making necessary for leveraging an entire business. They also reduce the cost of getting at information in a DBMS." The goals and vision of the governor, mayor or agency head must be tied to the data warehouse strategy. A government's information service agencies have to work with key decision-makers to ensure that their enterprise goals are linked to the architecture of the system, said Strange. Otherwise the data warehouse will fail, leaving users with nothing more than a series of stovepipe decision-support systems. To get an enterprise data warehouse off the ground, implementation should be done in stages. The first stage includes one or two decision-support applications, some ad hoc query users and a couple of subject areas of data. Subsequent stages increase the number of decision-support applications and subject areas supported in the warehouse database. Deploying a fully functional data warehouse is not only difficult to execute at the enterprise level, it's also expensive. GartnerGroup estimates that data warehouse costs can exceed $10 million, making it important to justify the risks involved against the benefits gained. Not surprisingly, out of approximately 2,000 companies that say they are planning to implement a data warehouse over the next two years, GartnerGroup estimates that less than 150 will actually complete the job. Despite the difficulties, data warehousing is becoming a mainstream activity as a growing cadre of large organizations turn to the technology for information access and decision-support. A number of trends indicate that the market is growing and will provide customers with a maturing technology over the course of the next several years. Some of these trends include increased support for data warehouses by leading relational database management system vendors, a shift by warehouse tool vendors to providing comprehensive solutions, and an increase in the warehouse consulting business, with systems integrators* bringing expertise and other advantages to the forefront. How states and localities regard the use of data warehouse technology depends on their vision of turning data into an asset. The Commonwealth of Massachusetts sees its data warehouse as a foundation for future uses of information, including the support of new interagency applications requiring shared data, better relationships with business partners and the private sector through electronic commerce and continued improvement in fiscal management. That kind of strategic vision should bode well for any government.
----------Sidebar----------
Putting a Warehouse on a Platform For its data warehouse, the Commonwealth of Massachusetts uses a network of symmetric* multiprocessing servers running several relational databases* that contain data from a number of systems that have been standardized to conform to common data definitions. The warehouse servers are connected via the commonwealth's wide area network to local area networks of PCs across the state. Data from the warehouse is delivered to the desktop using Windows-based query and reporting tools and executive information systems developed to provide senior managers with up-to-date information at a keystroke.
