Random Thoughts of a Scatterbrain.
 Tuesday, October 11, 2005

BizTalk Vs. Windows Workflow Foundation

10/11/2005 9:08:26 AM (Eastern Daylight Time, UTC-04:00)

Having worked with BizTalk Server 2004, SQL Server Integration Services (replacement for DTS), and having taken a quick glance at Windows Workflow Foundation, I was kind of confused why Microsoft would create so many competing technologies, especially considering the high license costs of BizTalk Server.

As an aside, during my time with BizTalk, I was thoroughly unimpressed.  To begin with, I hate working with technologies where you can't easily see every piece of the whole.  Too much of the functionality of BizTalk is hidden in property menus and context menus...I can't stand that type of crap.  Even more annoying than that is the dependency on the GAC.  It takes a ridiculous amount of time to build and deploy the assemblies.  Consequently, it's a PITA to test and debug.  I've not found any real scenarios yet where I've felt that I would recommend BizTalk over some other solution, especially considering the insane licensing costs for what amounts to nothing more than an XSLT engine with some workflow and connectors built into it.  People, especially Microsoft, will try to convince customers that it's the solution to everything.  They tried to convince one of our clients to use it for a task that was 10x easier and more efficient (time wise) to do using DTS.  Not to mention the client already had a license for DTS whereas it would have cost them in the neighborhood of $40k just to get a BizTalk server up and running. I dunno, maybe I'll be impressed by the next version.

If you're still interested, there's some discussion about the main differences between Windows Workflow Foundation and BizTalk on the web today, so I'd thought I'd share with anyone else that's interested:

The short answer is that WWF (no, not the one with big men in spandex) is a framework to be used to develop intra-application workflow whereas BizTalk is aimed at inter-application workflow and [buzzword alert] "business process management".

WWF looks promising.  Without realizing it, I actually built a mini workflow framework for a recent project I worked on which required automating the UI of Microsoft Project Pro 2003.  It allowed developers to program against an API to create "Steps" and "Actions" whose results could be linked to and iterated over by subsequent Steps and Actions.  Very cool stuff.

10/13/2005 1:31:30 AM (Eastern Daylight Time, UTC-04:00)
Please, get your facts right. BizTalk costs from as little as $1000.
mcgeeky
10/13/2005 8:30:57 AM (Eastern Daylight Time, UTC-04:00)
mcgeeky, I linked the page for the licensing costs in my post.

From the licensing page:

"BizTalk Server 2004 Standard and Partner Editions are intended for small-to-medium deployments, and are limited to a single CPU. BizTalk Server 2004 Standard Edition supports integration with up to 10 applications and 20 trading partners. BizTalk Server 2004 Partner Edition supports integration with up to three applications and three trading partners."

Yes, if you can live with those limitations, then the $7k or $1k version is just fine (but you still have to factor in costs for deployment).

Realistically, most customers will end up purchasing the Enterprise Edition, which is listed at $25k. This is not to mention the hardware costs, another 2-5k, the setup costs, and the consulting and/or training costs. So the initial deployment costs alone are probably in the $40-50k range. This is not to mention the cost of bringing in a BizTalk architect ($150-200/hr) and consultants to implement the architecture. I think you're looking at a $100-200k cost of entry, and even at the pricepoint, what have you accomplished that couldn't be done with some decent programmers?

I almost forgot, you also have to factor in a license for SQL Server 2000, since it's a requirement for BizTalk. So if, let's say, your shop runs Oracle or DB2, that's another $5-20k just for the SQL Server license.

I dunno. I guess I don't see the light yet with BizTalk. It seems like a lot of fluff and hype. So my point is, then, I don't see the _value_ that you get from your investment.

Admittedly, I have not worked on a large scale, real world BizTalk project, so I can't say for sure. But if you have more details on why you think that there's value in BizTalk, please feel free to expand on it beyond a two sentence comment ;)
Name
E-mail
Home page

Comment (HTML not allowed)  
Enter the code shown (prevents robots):

RSS 2.0 Atom 1.0 CDF