Discussions

News: Insight required: A viable alternative to enterprise Portal software?

  1. Enterprise Java customers in the portal space are starting to think twice about their portal software investments, and then after that, they're finding themselves thinking again. And it's the upgrade path from WebSphere Portal Server (WPS) 6.1.5 to WPS 7.x that's forcing them to search their souls.

    Rip and replace, or just replace?

    WebSphere Portal 7 doesn’t provide a simple upgrade path. Moving from 6.1.5 to 7.x means a fairly rigorous rip and replace procedure, and as corporations size up the migration to WebSphere Portal 7, they’re simultaneously asking themselves if this might be the right time to move away from the use of enterprise Java portal software entirely. But there’s an inherent problem in this proposition, owing largely to the fact that in the enterprise Java space, there simply isn’t a viable alternative to the enterprise portal.

    A legacy of broken promises

    Historically speaking, portal server software has never really lived up to its promise. Integration and collaboration were always big buzz worlds during sales calls, but in reality, the intensively personalized experience that Zuckerberg delivered through Facebook was something IBM always promised corporate clients, but at the same time, was something that never actually got delivered. Large corporate clients are now looking back at a legacy of investing millions of dollars into software that never lived up to the promise, and they’re wondering if moving forward in the exact same direction, with the exact same vendor, and the exact same set of licensing costs, is the right move to make. The move away from an enterprise portal is a concept that is eagerly embraced by development teams that have been burned by the portal development process, but the discussion at higher levels quickly turns cold when an attempt is made to conjure up alternatives. 

    Portal software: more than just branding and content management

    Any modern content management system worth its salt can generate stylized pages that can brand a website while delivering content across the world, so the whole argument of a portal being useful because of the way it renders skins and themes and does a little bit of content management is pretty much moot. Any web content management systems can do that.

    And of course, development with the Portlet API is hardly de rigueur. Much more rigueur would be the use of a lightweight client that leverages JavaScript, Ajax based request and RESTful web services in the backend so that Portlets and Servlets can be bypassed altogether. So in terms of content management, branding and lightweight client development, portal certainly doesn’t have the market cornered.

    How do you aggregate content without a portal?

    But there is still the task of bringing together these disparate types of enterprise applications and creating a unified, dashboard type of experience for end users. The portal starts to shine when layouts need to be easily managed and maintained, and integrated applications need to be easily upgraded.

    So, the question remains: “how do you aggregate and assemble applications together to provide a unified, dashboard type of experience to users, in the Java space, without leveraging the use of enterprise portal technology.”

    Such a task would be simple if there were only a few pages to manage; after all, Tiles, Facelets or even an exhausting use of JSP @includes could be used to bring together such an experience. But when you’ve got hundreds, if not thousands of pages being delivered to hundreds of groups across hundreds of departments, Tiles or Facelets simply isn’t going to cut it.

    Does a Java based enterprise portal solution ever become compelling?

    And this is where the portal argument starts to become compelling once again, because when you get stuck in the mud of page aggregation and application integration, a feature that few other technologies can do as well as portal, all of the other features of the portal, such as the ability to secure individual portlets, integrate with workflows, provide customization and perhaps  even more importantly, use vendor provided software that provides a whack of collaborative portlets  that can plug right into the system and be used immediately, all of a sudden you start to struggle with the question of what a large scale enterprise would base their architecture upon other than a globally deployed, enterprise portal server based product. God bless Nuxeo and Alfresco and  Magnolia, but as great as these content management systems are, they don’t compete with the feature set of a portal server like WebSphere, eXo or Liferay.

    Even enterprises that are questioning their future commitment to WebSphere Portal Server concede that few other software platforms provide the same promise and vision that a portal product provides. But when four or five years of sunk costs in terms of licensing hasn’t seen any of the true portal vision come to fruition, corporations don’t want to get locked into four or five more years of writing a vendor million dollar checks.

    Is the real solution an open source portal like Liferay or eXo?

    There is some great open source software currently available that rivals some of the big names in the industry. Everyone is impressed after the first time they install and play around with Liferay or eXo. Unfortunately, the term ‘portal’ has too often become a bad word, and when an architecture review board decides to move away from WebSphere Portal, they’re often wont to move away from the portal platform entirely. It’s difficult to convince someone who is dissatisfied with WebSphere Portal Server that their experience with Liferay is going to be any different. Plus, there is always an inherent  resistance many large corporations still have when it comes to embracing open source software, even when there is a large and stable company that stands behind the open source software in question.

    TSS readership insight requested: how do you wean a large, global, enterprise client off a heavyweight Java portal?

    Perhaps the TSS community can provide some elucidation on the topic. Is there an obvious alternative to using an enterprise portal, something that exists somewhere between installing a behemoth like WebSphere Portal Server on the left, and aggregating enterprise wide content using Facelets and a million little outsourced developers on the right? Enterprise architects are looking for a way out of paying Fifth Avenue license fees for enterprise portal software that has only been providing a Dollar Store experience to end users; and they don’t want to submerge themselves into another five or ten years of cost commitments without an escape hatch. The problem is compelling, but without an alternative to enterprise portal software, there’s really not much else that an enterprise architect can recommend.

    Further food for thought: Every enterprise needs an employee portal. Or do they?

  2. In an attempt to garner some community insight on the topic, I've reached out to both the Stackoverflow crowd and the CodeRanch crowd. Since those posts point here, it's only fair that this post cross-links to those threads as well:

    What is the alternative to a Portal Server Strategy - Stackoverflow

    How to move away from WebSphere Portal technology and architect a new solution?

  3. I wrote a small article about what technologies you will need to use instead of doing things with Enterprise Portal:

    http://lofidewanto.blogspot.de/2011/04/inside-out-and-outside-in-integration.html

    -> Outside-In-Integration:

    1. REST

    2. OAuth2 and OpenID Connect

    3. OpenSocial

    4. AppStore

    Cheers,

    Lofi.

  4. Maybe try mashups?[ Go to top ]

    About a year ago, over at SearchSOA.com we were hearing a lot form Mike Ogrins about using REST APIs to replace enterprise portals. His approach was to build a solid set of basic APIs that could then be mixed and mashed to generate business apps as needed.

    I'd love to hear what Java developers think about mashups now. Consensus?

  5. Geat summary[ Go to top ]

    This is the kind of posting I'd like to see on TSS again.

  6. /.[ Go to top ]

    Maybe you should ask slashdot too?

  7. What about Backbase?  Gartner says they are a solid alternative to traditional portals with their RESTful, WOA, widget architecture.  I like the way they loosely couple the presentation layer from the server, enabling the UI/UX folks to do whatever they like, while simultaneously offering several options for integration with core and 3rd party apps.  And, they sell to banks - who no doubt demand great security, scalability, etc.

  8. It would also be key to keep in mind how "Mobile" the solution needs to be. Gadget based lightweight alternatives provide some easy to implement options.