Enterprise Information Portal Architecture and Requirements

An Enterprise Information Portal (EIP) provides business users with a single Web interface to corporate information scattered throughout the enterprise. In this section we discuss the architecture of an EIP and review EIP product requirements.

Components of an EIP

The main components of an EIP are the business information directory, the publishing and subscription facilities, and the information assistant.

The Main Components of an EIP

Business Information Directory and Publishing Facility

The business information directory(BID) is a server-based index of an organization’s business information. This index is maintained via a Web-based publishing facility, by so-called meta data crawlers that regularly scan selected servers for new business information, or by an import interface that enables users and third-party vendors to maintain directory information via flat files or a programmic interface. The business information directory not only indexes business information like data warehouse relational tables, word processing documents, and Web pages, but also the decision processing objects (queries or analyses, for example) that are used to produce the business information. Users can employ the portal to find these objects, and, if authorized, run them. The business information directory not only enables an organization to document data (really meta data) about a wide variety of business information and where it resides in the enterprise, but also the ability to organize this information by subject area and topic. The directory in addition allows business users to annotate directory entries with additional meta data about the meaning and context of business information, and about the business actions that have been taken based on the information.

Subscription Facility

The subscription facility is used to control how business information viewed through the portal is distributed to business users. Information can be delivered or decision processing objects run – immediately, at a certain time and date, or at user-defined intervals. Business rules can be defined so that as information in a data warehouse database, for example, changes the rules are evaluated and if satisfied (a client’s total stock portfolio reaches a certain threshold, for example) a report generated automatically and delivered to the user. The subscription facility supports not only explicit subscriptions to information, but also implicit ones. An example of an implicit subscription would be if the user defines to the portal (most probably in a user profile) interest in certain types of information, or if the user belongs to a group of users associated with certain types of information. When a piece of information is published to the business information directory, the portal distributes automatically copies of the information to users whose profiles indicate an interest in the type of information that is being published. Another form of implicit subscription is unsolicited information where a user may decide that other people in the organization need to see a copy of a particular piece of information, and a request is made to the portal to deliver the information to those users, either on a one-off basis, or at regular intervals.

Information Assistant

The information assistant provides a fully customizable Web interface that works in conjunction with a search engine to enter and process user requests for business information.

EIP Product Requirements

Outlined below are seven key requirements that an EIP should support.

1. A portal should provide a business information directory (BID) for maintaining meta data about an organization’s business information. This meta data should not only contain details about the business information itself, but also about how the business information is related to other information objects in the business information supply chain. Users of the directory should be able to drill-thru from the business information directory to technical meta data maintained by a meta data interchange hub and its associated technical directory. More information about a meta data interchange hub can be found in Managing Distributed Meta Data. The ability to define meta data relationships within the business information directory, and between the business information directory and the meta data interchange hub enables the complete flow of information from transactional processing systems to decision and collaborative processing systems to be documented and tracked by an organization.

2. The business information directory should have an administration capability for defining user and user group profiles. These profiles should allow the administrator to define and control the types of information a user may see, and the features of the portal the user may employ. Users should also be able to modify their personal profiles to indicate the types of information they are, or are not, interested in receiving.

3. A publishing facility should be provided that enables both technical and business users to employ an interactive Web interface to document the business information that exists in an organization. This facility should support a wide variety of different information types including relational database tables, multidimensional databases, word-processing documents, spreadsheets, images, video, audio, HTML and XML pages, e-mail messages, and decision processing objects such as queries, reports, and analyses. When publishing information, the user should be given the option to move the information being published to a shared and protected information store. The user should also have the ability to annotate business information directory entries with additional information.

4. The publishing facility should provide crawlers that automatically scan/index the contents of specified file directories for new business information. The frequency of scanning, the directories scanned, and the file types processed should be controllable by the administrator. The results of the scanning/indexing process should be placed in a staging area for editing and entry into the publishing facility by support staff. The scanning/indexing process should support installation-defined exits that can be coded in industry standard programming languages.

5. The publishing facility should support file import and export facilities (ideally via XML), and a documented programmatic interface that allows external products to access and maintain meta data in the directory.

6. A subscription facility should be provided that allows business users to have information delivered to them on a regular basis. This facility should allow the user to control via a Web interface the delivery schedule, the format the information is to be delivered in, and the location to which the information is to be delivered. Scheduling options should include immediate delivery, a specific date and time, a regular schedule, and event-driven triggers controlled by user-defined business rules. Delivery formats should include popular Web and office-system file formats. Destinations supported should include Web addresses, e-mail user-IDs, pagers, and fax machines. The subscription facility should allow the user to schedule the running of decision processing objects such as queries, reports, and analyses created by leading business intelligence tools. It should also support implicit subscriptions whereby information is delivered automatically to users based on their profile, or when requested by another member of their user group.

7. A simple and easy-to-use Web browser information assistant should be provided for accessing, maintaining, and navigating the business information directory. The information assistant should have a search capability that allows the business user to scan directory contents based on user-entered search criteria. The information assistant should provide different styles of user interface that can be tailored to suit the needs of different users and the type of work being performed. Possible interfaces include the traditional desktop file folder interface, Web search engine interface, and the hyperbolic tree. The Web search engine interface is more suitable for inexperienced users, whereas the hyperbolic tree can display more information and is more suited to experienced users who wish to navigate and drill-down through information.

Choosing the Right Product

When choosing an enterprise information portal, the list of requirements presented in the brief should be used to assess user needs and to evaluate products. The main distinguishing factors between products are likely to be the power and openness of the business information directory, and the ability of the product to support a large number of users and a wide range of business information. These factors are important for organizations that wish to create an enterprise-wide EIP strategy for organizing and finding business information. Without such a strategy, it is likely that multiple portals will be deployed, which will lead to significant information integration problems, and defeats the key objective of an enterprise information portal – providing business users with a single interface to business information.