Portal products lend themselves to rapid deployment. Use of prebuilt integration components dramatically affects the resources and effort of deploying the 1.0 release,
compared with subsequent releases of an enterprise portal.
Key Issue
Strategic Planning Assumption
Many portal product vendors provide good "out-of-the-box" functionality that is, the vendor offers a long list of off-the-shelf integration components (e.g., portlets, gadgets, OMlets, Iviews, Eclips) for integrating a variety of applications and repositories. This functionality delivers to the end user a rapid prototype and fast deployment of the initial release of an enterprise portal. Nevertheless, out-of-the-box functionality is only indirectly tied to the overall architecture of the portal product. It is, rather, a manifestation of the partnerships created by the portal product vendor and the effort invested in providing this method of integration.
New Rule of Thumb. For the initial release of an enterprise portal (release 1.0), 90 percent of the capability will be delivered by the out-of-the-box features of the portal product, while only 10 percent will be delivered by custom integration components.
For subsequent releases of an enterprise portal (releases 2.0, 3.0, etc.), this ratio reverses, with 90 percent of new capability delivered by custom integration components, while only 10 percent of new functionality will be delivered by new
out-of-the-box features.
This 90/10-10/90 rule has dramatic implications for the total cost of ownership (TCO) of the portal, the skill sets required, timing for subsequent releases, maintenance to back-end applications and overall architecture of the portal product.
TCO. Although many enterprises think that the major costs of the portal go into the product acquisition and initial release, those that have done their homework recognize that the long-term TCO of the portal is driven more by subsequent
customization work.
Custom development of integration components is required for all those integration points that are not satisfied by the off-the-shelf portal integration components (portlets) or middleware/enterprise application integration (EAI) capabilities. This custom development can be extensive, and must be completed by internal developers or outsourced to consultants or systems integrators (SIs) so these costs can be tremendous. Gartner encourages clients to budget one to three times the acquisition cost of the portal product for development, deployment, maintenance and support of the portal during its first two years of use.
Skill Sets. Custom development of integration components requires highly skilled developers. Those that follow the integration component approach must write code that calls for low-level application programming interfaces (APIs) or SQL statements. This code must be implemented in a variety of programming languages, depending on the portal product vendor selected. The majority of portal product vendors support Java for this purpose and skilled Java resources are expensive, and will remain in short supply through at least 2002.
Timing for Subsequent Releases. The 1.0 production release of the enterprise portal usually occurs within four to six months of initiation of work. The 2.0 release, which requires significantly more custom integration development, may take much longer to deliver. It's a mistake to try to deliver additional functionality in a single release rather than through point releases. Iterative approaches should be used in the initial release, as well as in subsequent releases.
Maintenance to Back-End Applications. When back-end applications are upgraded or changed, the integration mechanisms for those applications must also change. If the integration mechanism is a portlet, these updates and changes present opportunities for software to get out of sync. The portlet model can be somewhat brittle changes in the back end frequently break the portlet. The effort of coordinating updates to both the portal and the back-end applications will be a drain on enterprise resources. Because most portal products don't have sophisticated change management mechanisms, this synchronization will have to be done manually.
Overall Architectural Requirements. Portal products must have two major application integration mechanisms:
Using an abstraction approach for application integration is much more efficient than writing new portal integration components. Additionally, correlation across stovepipe applications including process integration is much more feasible through abstraction approaches, as data and processes that are abstracted are more easily tied together.
Enterprises should 1) look for portal products that contain both methods of application integration and 2) recognize that although a long list of off-the-shelf integration components can deliver tremendous out-of-the-box functionality, exclusive use of this integration method will be more costly to maintain over time than using abstraction approaches.
Bottom Line: Enterprises should plan to spend one to three times the acquisition cost of the portal on the development, deployment, maintenance and support of the portal during a two-year time frame. Through 2004, immediate functionality (via portlets) and abstraction mechanisms will be equally important in portal product selection. Enterprises should plan for a major shift in resources and timing when the enterprise evolves from r.1.0 to subsequent releases, and should plan for significant extra effort as back-end applications are upgraded.
Source: Gartner's Internet Strategies Research Note Strategic Planning, SPA-14-5861, 3 October 2001, G. Phifer.
Front Page
A Seismic Shift in Business, A Perfect Storm in Software
Bowstreet Business Web Factory Data Sheet
Gartner Files: The Future of Web Services: Dynamic Business Webs
Gartner Files: Service-Oriented Development of Applications: SODA Pops
Gartner Files: The 90/10-10/90 Rule for Portal Deployment
Gartner Files: Bowstreet Refocuses as Web Services Become Mainstream
Bowstreet Web Services Report Web Letter is published by Bowstreet. Additional editorial material supplied by Gartner, Inc. © 2001. Editorial supplied by Bowstreet is independent of Gartner analysis and in no way should this information be construed as a Gartner endorsement of Bowstreet's products and services. Entire contents © 2001 by Gartner, Inc. All rights reserved. Reproduction of this publication in any form without prior written permission is forbidden. The information contained herein has been obtained from sources believed to be reliable. Gartner disclaims all warranties as to the accuracy, completeness or adequacy of such information. Gartner shall have no liability for errors, omissions or inadequacies in the information contained herein or for interpretations thereof. The reader assumes sole responsibility for the selection of these materials to achieve its intended results. The opinions expressed herein are subject to change without notice.
|