Flex Data Services is an interesting product slowly finding its way into the market place though I have to admit I haven't gotten many project proposals in that want to go down this route yet. There are obvious benefits to the technology particularly for those coming from a Java background and starting to use Flex (just look at the support for the Hibernate and Spring frameworks). We're probably seeing the most Flex adoption coming from that side of the spectrum now, where a lot of Flash developers were already sold to the concept early on but had financial constraints prior to Flex 2 to work within.

We can't complain about pricing nowadays with the Flex 2 SDK and FDS Express being available free of charge, yet there are still a few obstacles that prevent it in my opinion from going mainstream any time soon.

FDS Express Its not clear how far FDS Express will get you, the license restricts you to a single CPU so you probably won't be able to run something that needs a lot of processing power. I suppose it should do just fine for limited intranet use, but will it prove to be usable for a public facing application?

Obviously Adobe wants to sell licenses but the issue there is (as with a couple of other products) that there is no good sliding scale pricing, with the cheapest edition (after FDS Express obviously) coming in at 6,000 USD. I believe there is a huge untapped pool of Flex developers out there that would fork out a couple of hundred dollars for an entry level license of the product.

The overall problem might be with the way Adobe wants to position this as an enterprise product, and if its pricing starts under a couple of grand it isn't taken seriously by a lot of corporations.

Documentation There is quite a significant learning curve getting started with FDS, particularly in setting up and configuring it for your application and although it has improved fairly little documentation is available from Adobe or the community.

Lack of a killer feature Those of us coming from a Flash development background often find no compelling "killer feature" in FDS. I know at least that was my reaction when I first saw it demo'd, the only wow factor was how few lines of code it took. Any combination of a socket server and Flash Remoting gateway (and a lot of elbow grease) could roughly speaking do just about the same.

I know Ted Patrick built a wonderful AMF server using BinarySockets in AS3 a while back and I'd venture to guess that we'll see similar OSFlash initiatives emerge in the coming year.

Heck, I'd be interested to work on that if Patrick doesn't already have it on the roadmap for AMFPHP -- most of the work would be in doing API's for messaging, queuing and conflict handling, as the hard lifting has already largely been done by other projects.

My suggestion for Adobe would be to have Flex Data Services go the enterprise route with strong support for Java and .NET (?) and reviving Flash Remoting with push support for the rest of us. At least that way we'll see both ends of the market covered with a product specifically aimed at their needs.