A public forum for discussing the design of software, from the user interface to the code architecture. Now closed.
One customer wants three of their applications to exchange data: ERP, warehouse system and manufacturing system. We just know what should be exchanged, but do not know how (we will probably need to ask the application providers to tell us about the tables on the databases or to provide some kind of XML interface)
My boss told me to evaluate the costs and needed resources using Microsoft Biztalk. Since I have never used Biztalk and this is my first estimation I am feeling kindof lost...
Any help will be really appreciated. Thanks.
Monday, May 22, 2006
+1, it's the most powerfull yet underestimated piece of software creted by MS.
Go for it.
Monday, May 22, 2006
I have found BizTalk to be too bloated for the real world. When your transform mappings get too big, you have to stop using the visual designer because it just croaks. On the other hand, you don't need the complexity of BizTalk for the smaller jobs, so why even bother?
Also, error handling is a nightmare. Sure, it all makes sense when you first put it together, but try following the logic of that crappy diagram 9 months from now when production is down and your primary BizTalk developer is on vacation.
Another poster recommended training classes - I second that ---- for your _whole_ team. And make sure your whole team is involved in the design and development of all your BizTalk solutions. It's almost impossible to open a BizTalk solution you have never seen and fix a problem in a timely manner.
I haven't used the current version of BizTalk, so things might be a little better now.
I have recently implemented RosettaNet Accelerator 3.0 with BizTalk Server 2004. The accelerator was pretty buggy for the 5K we spent on it. The support available for it was negligable. But it is Microsoft so it passes the corporate standard. Biztalk itself has some nice features. My biggest complaint was how complicated the install and configuration is. I am really pleased to see that all that complexity seems to have been eliminated in BizTalk 2006. I have yet to evaluate the RosettaNet accelerator 3.3 that comes with Biztalk 2006. I am hoping they have fixed the bugs in the 3.0 version. I really like the licensing changes for Biztalk 2006 standard edition which fits us perfectly.
In summary Biztalk is worth a look if you have .net skills. But the ironic thing about it is if you have moderate .net skills you could easily knock off custom programs to do the same things.
Just to clarify..
5K for Biztalk
5K for the RosettaNet Accelerator
Current Licensing from Microsoft charges around 9 or 10K for Biztalk (standard) but includes all Accelerators. And they redefine the license so you can have unlimited trading partners on 2 cpu's but limited to 5 applications.
A dedicated server could handle a lot of traffic with this license model.
> It's almost impossible to open a BizTalk solution
> you have never seen and fix a problem in a timely manner.
Yup. Your whole team should get training. Even though from BizTalk 2004 onwards Microsoft replaced the original abortion of a designer based on Visio, the current designer still isn't something you'll pick up in a hurry; its scope is too broad.
I also agree with some other posters; solutions aren't any easier to develop on BizTalk's J2EE alternatives.
+1 for BizTalk (preferably 2006 for it's SUBSTANTIAL improvements in deployment / ease of use).
1. It may get very VERY expensive very VERY quickly if you need multiple server (enterprise version license retails for 24K / cpu).
2. The learning curve will be steep. You probably have all the tools (skills) you need to craft a solution, but the paradigm is different.
3. +1 for sending your whole team to training, but find a good trainer. I had a client select a training company (alas, without my input ... could've saved 'em a ton) but the class was a disaster. Instructor read from a book and made no references to real world solutions deployed all around him ... students walked out and went back to their real jobs.
4. BizTalk is a HUGE product ... it is likely that your solutions will use only about 1/3 of total (functional) capability. Don't worry about that ...
I think there's probably more, but brain hurts at moment for reasons unrelated to BizTalk.
Wednesday, May 24, 2006
I know what I going to say (slaps self in forehead) ... if you need to deploy this solution next week or next month, look elsewhere. If you have the luxury of time, sounds like a pretty good fit.
Wednesday, May 24, 2006
Biztalk is a very fine product for these types of applications. I have successfully designd similar setups with it in the past. Pricewise it beats the hell out of the competition (try getting a quote for SAP XI (basically the same type of sofware) and be prepare to be unpleasantly surprised).
If the customer can't afford a BTS licence, make sure you get a good T&M contract. Yes you can cobble up a DIY solution on the cheap (look at Altova tools for a fast generate 80% of the solution aproach), but I would advice against this for business critical intergration projects.
Just me (Sir to you)
Thursday, May 25, 2006
This topic is archived. No further replies will be accepted.Other recent topics
Powered by FogBugz