Most senior executives, across all enterprises, are likely to agree that two of their most valuable corporate assets are their employees and their corporate data. Most will also agree that increasingly these employees need better, faster and more complete access to information in support of their daily activities, so that they can better contribute to the efficiency and profitability of the organization.
Yet enabling these employees to gain easy access to corporate data is still something that many of today's businesses have failed to achieve. The Gartner Group has described this issue as the "Fact Gap," and a recent survey by the research group Find/SVP revealed that 84 percent of IT executives in large corporations have been mandated to close this gap by finding ways to provide information to workers who need it, when they need it.
The Evolution of Information Reporting Solutions
Why does this "information gap" exist? An examination of the evolution of reporting solutions gives some understanding of the issues that organizations have faced in trying to deliver information to their demanding user populations.
In the 1970s data was stored in mainframe databases which were hard for end users to access. They were forced to rely on IT personnel, who used programming languages, such as COBOL, to run, print and distribute their reports. This provided little, if any, flexibility to end users who needed information immediately or whose information needs changed rapidly. The emergence of 4GL programming languages improved the speed with which IT could respond to demands for customized information, but use of these tools rarely reached beyond the programmers. The result was an ever-growing backlog of end-user report requests.
Growing availability of personal computers dramatically changed the reporting environment, and an abundance of desktop tools were delivered beginning in the late 1980s. These tools provided end users with better access to data; and for the first time, some end users were able to create their own reports. However, these tools were not the hoped-for panacea. Typically they provided only simplistic access to corporate databases and were unable to cope with the real-life complexity of an organization's IT architecture or end-users real information needs.
These tools, now often described as "fat clients," also placed unforeseen stresses on the IT department. Networks became overloaded by end users attempting to download large data sets from the mainframe to their PCs, and the demands on IT for support and training increased significantly. By the early 1990s, the reporting market was firmly segmented into two types of tools: end-user-oriented desktop tools and programmer-centric production reporting tools.
More recently data warehouse and data mart implementations have made corporate data more available and useful. Additionally, the Internet and corporate intranets have emerged as the preferred means to deliver information because they are easy to use, fast to implement and cost effective. Now any user with a standard Web browser has an easy way to view information--as a document, with embedded hyperlinks to other documents. This new content-based paradigm allows users to work with information in a more understandable and comfortable way and sets new expectations for future reporting solutions.
The Enterprise Reporting Solution
The META Group has suggested that a new class of software, enterprise reporting tools, will rapidly emerge during the late 1990s and overtake the deployment of both desktop and production tools. This class of tool combines some of the best attributes of the desktop and production reporting tools with a three-tier architecture to provide a range of functionality and scalability capable of supporting the large deployment needs of the enterprise.
The META Group has determined that the information needs of an organization can be broken into four user categories, as in Figure 1.
The information needs of each type of user determine the type of reporting and analysis tool and the appropriate environment (Web or client/server) an organization should use to meet the needs of the user. As a rule of thumb, the less interaction required, the more suitable a Web-based delivery mechanism will become. Conversely, the more interactive the usage need, the more appropriate a client/server environment becomes.
Enterprise Reporting Emerges
The user definitions described earlier demand a new breed of enterprise reporting tool to support the broad range of users and their information needs. This new breed provides different user versions of tools which deliver the appropriate level of functionality in the preferred environment and which incorporate a powerful middle-tier report server component. This three-tier model (with client/desktops and database servers comprising the other two tiers) allows for server processing and the large-scale information dissemination necessary to support large numbers of users.
The new breed of enterprise reporting tools combine a number of key characteristics, as indicated in Figure 2. They take advantage of today's easier data access; leverage the content-based paradigm popularized by the Internet, intranets and the Web; and give every type of user a level of functionality suitable to their needs.
Enterprise reporting tools should support the myriad of information needs of a typical organization by providing functionality suitable for different types of users.
A trained user, power user or IT professional will be able to quickly produce polished presentations of information that contain data combined from multiple queries or even multiple data sources. By making all the contents of the report document "active" or controllable, large volumes of data can be turned into a series of documents which are hyperlinked in a logical hierarchy. With zero training, report consumers begin with a "content-rich" document and work in a comfortable way with their information. Viewing and navigating from summary to detail reports is a content-driven process. Delivery of these "live" reports meets the information consumption needs of the broadest class of users.
To facilitate light analysis, the tools must provide simple ad hoc query capabilities. To do so requires a meta data layer, or repository, that describes the underlying warehouses or operational databases in business terms. The user does not have to wrestle with obtuse table and column names or have to formulate "join" strategies within their query or report--all of these are handled within the repository itself. The repository will also manage other valuable elements, such as common calculations. The more flexible repositories will interact with existing meta data to smooth set-up and maintenance tasks.
Heavy analysis needs should be met with appropriate data exploration tools. Working with specialized data marts or data hypercubes, these tools meet the needs of those users who need highly interactive analysis of large, generally summarized data sets.
Dissemination of reports should be automated and managed from the report server. Management tasks may include notifying users that output is ready, automatic printing or publishing to HTML for Web browser users to access and sending the output to e-mail distribution lists. Integration with e-mail is critical and allows many users to receive and reuse information across the organization and beyond.
There is no doubt that the rapid expansion of the Internet as an information dissemination mechanism will have a major impact on enterprise reporting solutions. The arguments for moving to this method of deployment are compelling. Organizations are insisting on the ability to "build once, use anywhere" and demand enterprise reporting tools that support that effort. This means that a report built and designed for client-side viewing today will need to be available as an HTML page for browser viewing tomorrow--without having to be amended. Enterprise reporting solutions are ideally positioned to solve both requirements. Their server-centric architecture is ideal for leveraging the Web. And their ability to allow organizations to seamlessly deploy the same report to both client/server viewers and Web browsers ensures that organizations can achieve the best of both worlds--a truly enterprise-wide solution.
One of the biggest issues faced by adopters of desktop-centric tools was their inability to scale beyond small user communities. Enterprise reporting solutions solve this by utilizing larger middle-tier processors, typically running under UNIX or Windows NT. Reports can be submitted to the report server, run and then be easily distributed to a large number of users without each user downloading large data sets and then performing further processing on the desktop. By completing 100 percent of the processing on the report server, including all formatting, an enterprise reporting tool can dramatically increase the throughput of information.
Reports on the server can be run immediately--on-demand or scheduled to run at predefined intervals. Schedules can be based on time intervals or be calendar based (e.g., every Friday). Since immediacy of information has become critical, largely due to our increased expectations set by the availability of information on the Web, the capability to schedule even in minute intervals--rather than hours or days--is crucial. Reports may also be triggered by external events or have dependencies built in so that a report won't run unless a preceding task has completed.
A well-designed and truly scalable report server takes advantage of the strengths of each component in the IT infrastructure. Full exploitation of the server environment ensures greater reliability, availability and better administrative information such as error logging. Multithreading ensures that the most efficient use of each server is made. If multiple servers are a requirement, then load-balancing and fail-over facilities may become important. The enterprise reporting solution must also make the best use of the target database(s). Most reports and queries use some database functions, such as aggregate statements like GROUP BY or the application of filters. These should all be processed in the database, thus taking advantage of its power and indexing. Returning large amounts of data to the middle-tier report server (or worse, to the desktop) will greatly restrict the ability to deploy a successful solution. This factor becomes particularly important considering the very large databases found in many organizations today. The report server should also provide the automatic facilities to stage the data into appropriate hypercubes to support the needs of the "heavy analysts."
While it is important that users have the flexibility to help themselves to the information they need, it is equally important that an enterprise-wide solution provides manageability and control.
An administrator must be able to view what is happening on the report server, or servers, and quickly and easily address any issues that arise. This kind of control helps to ensure that all users continue to receive their information when they need it by effectively managing the large volume of reports and queries that the enterprise reporting infrastructure is able to handle. This will mean being able to centrally control aspects such as resource utilization and information security.
Enterprise reporting technology provides organizations with unprecedented flexibility when implementing enterprise reporting solutions. They are able to address the key issues left unsolved by both desktop and production reporting tools. Deploying such a solution will dramatically improve an organization's ability to close or even remove the "information gap
Charles R. (Rick) Chitty is president, CEO and co-founder of IQ Software Corporation. Before founding IQ Software in 1984, Chitty was with Datascan America (Online Software, Inc.) as secretary/treasurer-controller and director of support and international operations. He is a board member of the Southeastern Software Association and 1993 winner of the Ernst & Young "Entrepreneur of the Year" award.