Saturday, May 20, 2006

Has Flex given up on the enterprise?

One of the many benefits of adopting a SOA in the enterprise is the increased organizational agility that enterprises gain from a new type of desktop application that orchestrates the services exposed by the enterprise's legacy systems in a more joined-up way that better matches each business process.

This new type of application is called a composite application by IBM, Sun, SAP, BEA and many others including innovative ISVs such as Digital Harbor and Above All, who already strongly position their offerings as development environments for composite applications.

In reality Flex is the best tool I've seen for the rapid development of robust composite applications for the enterprise (and I include Windows.Forms in that) but Flex is also one of the industry's best kept secrets. There's no mention on Flex's Home Page of its capabilities as a development environment for composite applications. Instead Flex is positioned as a tool for developing RIAs. What are they? Websites with fancy graphics?

Right now IT departments in enterprises are pulling their hair out trying to work out how to deliver better smart clients to the business that compose the services exposed by their SOA and that can be easily deployed and reconfigured at low cost. "Flex can do that", I hear you say and you'd be right. The trouble is enterprises don't perceive themselves as building internet applications, rich or otherwise, so (wrongly) don't see Flex as the solution to their problem.

7 comments:

Anonymous said...

Weren't alot of Flex 1.x production applications internally deployed as enterprise apps anyway?

Maybe they're trying get at the Web 2.0 crowed with Flex 2

Anonymous said...

I don't really think peopoe are pulling their hair out over this. Just about every major vendor out there (IBM, Oracle, BEA, SoftwareAG, Sonic, etc. etc.) offers tools to build composite applications and automate business processes. The core to most of these offerings is an ESB, with process orchestration built on top of that.

I think you make a good point that Flex could be useful in this space, but in my mind, it still only plays a small role in the overall assembly of a composite application. In reality, Flex itself lacks too much of the critical infrastructure pieces to "do it all", at least when compared to what's already out there in the market. Perhaps it might give you process orchestration lite, but definitely not of a robust offering to satisfy what most enterprises are looking for.

I do see it providing a nice view layer for what the underlying app server has pieced together, though.

Anonymous said...

SAP has Flex embedded. And, yes, a lot of apps are deployed internally. Bi-weekly I'm writing answers to questions that Enterprises want to know. It's been really challenging to write in their language, but yes, some are interested, and they are getting it. Slowly, but surely.

Anonymous said...

I think the market will decide as usual. Adobe doesn't want to be a middleware company so if Flex 2.0 takes off in the enteprise for the UI stuff, then I'm pretty sure the IBMs, BEAs, Oracles, of this world might provide hooks into FDS etc.

Anonymous said...

For what it's worth, I was first exposed to the phrase "composite applications" and its workflow implications during Adobe orientation meetings soon after the acquisition closed. Since then I haven't seen many other references in weblog conversation, and so have held off myself.

Tying together various remote services in a customized local interface... makes sense, in a number of different workflows.

Flex awareness is on a rapid upswing, but it still isn't as widely recognized as it will be, agreed.

Anonymous said...

I concur. We have been using Flex 2.0 since Alpha and have found it an awesome platform for developing Enterprise Apps. Sure there are some missing elements particularly around its database management functionality (eg. things like no native transaction commitment control and database table access is somewhat clumsy) but nothing that can't be reasonably easily overcome (we have developed our own PHP-based ADODB libraries).

Another big clue to Adobe consciously or unconsciously missing this market segment is the lack of a category for Enterprise Applications in Adobe's Flex Developer Derby.

I think the lack of visible addressing of this market segment is a mistake. With Abobe's biggest threat, WPF, on the horizon, Adobe need to immediately harness their (current) key competitive advantage's which are its (relative) immediate availability, Flash Player ubiquity/mindset and cross-platform access. The division between Enterprise and whatever "RIA" means to the marketplace is going to grow increasingly hazy with the increasing uptake of SOA so ignoring the Enterpise segment is a risky strategy.

Come on Abobe - even though this is a new market for you, you really do have a winner on your hands.

Anonymous said...

Agree.

WPF and WPF/E is definitely a threat, especially with its platform XML UI language vs Flex's Application Framework XML UI language.

Adobe should really try to pitch Flex hard (and Apollo in the future) for the enterprise because its such an ideal product for that market.