Ah, Portal. It’s the answer to all of our content troubles: SSO; content aggregation; personalized pages; “easy” portlet integration; etc. There exists an extremely large number of companies that are getting raked over the coals for portal solutions that they may not even need. In fact, I personally have been involved with or worked around 3 portal projects that have had serious problems. Colleagues of mine have experienced the same issues with projects many times that, collectively. How is it that so many portals have such a high failure rate?

Buzzword: Portal

In many cases the answer simple - It’s a Buzzword. Company after company seems to be falling victim to the same sort of hype that got so many development teams bogged down with technologies like entity beans (see: To EJB or not to EJB) and the like a few years ago. Except this time it’s the Portal technologies. Admittedly, portals have their place. They really are good at aggregation of content of disparate systems and managing SSO. And personalization of portlets and pages on a “dashboard” can be a very compelling business proposition. There are, however, some caveats.

You will need a good business case that calls for the aggregation of content from many systems. If you don’t have multiple systems with unique data to be aggregated then you will want to stop and think really hard about how much you like wasting your time digging through the guts of a portal framework. If you simply need SSO, then the overhead of a portal really doesn’t give you a clean solution to your problem. The same would apply if you simply needed to aggregate a small set of data with no SSO requirements.

Implementing a portal (or anything else for that matter) without good reason is likely going to be the source of many problems. We’ve all been guilty at some point or another of implementing a technology (i.e. EJB) just because we wanted the experience or just because it was a hot topic. However, as we reflect on that we tend to see that a simpler (or different) solution was probably the more appropriate choice.

Portal Overhead

This is truly an important consideration. Anyone that has worked with some of the major portal vendors knows from experience that a portal is anything but easy when it comes to implementation, integration, customization, and deployment. If you indeed need a portal watch out because this is going to be one of the more brutal tests of your patience.

Someone new to portal development and implementation is likely to grossly underestimate the amount of time it takes to go from “We need a portal” to “The portal is complete”. Most of the portal implementations to which I am aware are supported by full-time staff. A team new to a portal implementation will likely start a project and then fail to deliver due to unforseen complications. Then a consultant (or I’ve seen whole companies hired) will come in and take the reigns and eventually after enough money is thrown at it, you will find yourself staring at something that looks like a portal.

A team experienced with a particular portal vendor will often be able to shave a lot of cost off of the delivery of a portal. But a company could find themselves spending many months even up to a year or more (and oh so much money) getting a portal out the door.

Portal Solutions

Portals are packaged sets of solutions. If you can’t truly say that you need a majority of the features provided by a portal then it’s probably safe to say that you don’t need a portal. Most are broken up into layers of “value-contributing” components. But they are all different in their own way. If you can say you need one, do your research before making any decisions. Manageability offers a good list of open source portal offerings out there. Alternatively, you may find some of the other commercial vendors have what you may be looking for.

The offerings and landscape are always changing and I hope that a trend toward ease of use is developing. The commercial Portals aren’t cheap. Not only are some portals expensive in terms of licensing, but the cost of ownership is much higher and should be analyzed and factored into the decision-making. Don’t use a portal for the fun of it because you will be seriously disappointed. If you do need to use a portal be sure you seek good objective advice and help in your endeavor.

Tags: antipatterns, architecture, java, patterns, portal