Sunday, October 23, 2005

Another false start for Microsoft's business apps

Microsoft just can't seem to finish anything related to business applications these days.

The latest case in point is the Microsoft Business Framework (MBF). Microsoft appears to have all but given up on building the MBF, sending half of the 200-plus development team back to Microsoft Business Solutions (MBS), the unit responsible for business applications, and half remaining with the Platform and Tools division, where they've been working for the past two years.

MBF was to be a set of lower level software application components to provide core business functionality upon which higher level applications and systems could be constructed. The original plan was for MBF to be the underlayer for Project Green, the successor to Microsoft's current disparate business applications, such as Great Plains, Axapta, Solomon, and Navision. MBF would be an open platform that would also allow third-party developers to build industry specific applications that would easily interoperate with Microsoft's own business apps.

Microsoft, of course, is claiming that the break up of the MBF development team really doesn't mean anything, and that MBF will still be delivered, but just in a different way, with bits and pieces showing up in other products, etc., etc. Mary Jo Foley at Microsoft Watch has a good summary of the history of fits and starts with MBF.

One Spectator reader has given me some interesting insights into the history of MBF and the current situation with Microsoft Business Solutions. Note that my editorial comments are in [brackets].

The pre-history of MBF originates from then independent Great Plains where they worked on the next generation Business Apps Platform. One of the reasons that Microsoft bought Great Plains was some promising ideas from the Great Plains R&D team.

However, the release of even bits of the new generation technology was characterized by constant delays and problems. At Stampede 2002 the MBS people even joked that every year since 1999 a new and better thing was promised but either was not delivered or when it was launched the result was a failure. Mostly it was related to portal technology and it was more related to Solomon than Great Plains.

The first working commercial piece [for MBF] was Business Portal 1.0 - now in C# sitting on top of MS Portal technologies....

Surprisingly, the [MBF] development environment for the entity persistence model was Rational Rose. Microsoft distributed a MBS Great Plains branded Rational Rose trial as THE development environment for MBF, but alas, IBM bought Rational Rose. Then the big issue was to tweak MS Visual Studio to accommodate all that was lost when they had to abandon use of Rational Rose.

There was a big undertaking by the MBF team and the Navision team in Denmark to port Axapta to .NET-MBF, but that effort was soon abandoned.

Part of delay for MBF may be associated with the delays and stripped down functionality for Longhorn [the next generation of MS operating systems], but I do not believe that this is the core reason. The most important reason for the delay of MBF is the delay of WinFS [the next generation Windows file system]. In the entity persistence model, the integrated security model is actually the core function related to MBF.
My source goes on to provide some insights into MBS product development in general.
In the meanwhile in 2002-2003, project Magellan was started....The outcome was MS Office Small Business Accounting, which was a total re-write in C# and MSDE. The database layer for this product is very heavy, with hundreds of stored procedures, functions, and triggers. It is really more like Great Plains, not like Axapta and Navision. The database layer is so heavy and the database normalization structure is so far thinking that it makes me think that this is the way the next generation unified business applications will be developed.

The packaged software thinking of Microsoft was reflected in many product policies over time. Now they have finally switched back to developing all products independently and trying to rescue Project Green by releasing it in waves.

Surprisingly, Navision is not a friend of MS SQL Server at all. There is big degradation at 30 users and a plateau at approximately 100 users. Furthermore, Axapta scales better on Oracle than on MS SQL Server, and it does not support SQL Server 2005 up till 4.0 or even later.
He also speculates on potential acquisitions and the situation among the MBS VARs.
It is possible that Microsoft will start buying market share. Epicor may be the prime candidate, but there may be others.

Just as there are problems with MBS being profitable, the MBS VARs also have been in trouble, especially the large ones. Some have been running at huge losses for several years before being bought by someone. At least so far this has not resulted in Microsoft buying some of its big VARs--that would definitely upset the community. But watch out for hidden support measures, such as more discriminatory margin structures or industry builder initiative financing.
Although this entire post paints a pretty bleak picture concerning Microsoft's development efforts in business software, I am not by any means suggesting that software buyers stay away from Microsoft's offerings. I continue to short list Microsoft's business solutions where they appear to be a good functional fit. Microsoft may be having difficulties, but it has the resources to deliver--something that can't be said for many of its competitors in the small and mid-sized business market these days. But first, it needs to get its act together and start delivering stuff.

If you have different information, or an opinion on anything in this post, please feel free to email me or use the comments section of this post.

Related posts
Reorg highlights troubles at Microsoft Business Solutions
Microsoft to put enterprise applications on the auction block?
Is Microsoft dying?
Microsoft wants PeopleSoft customers but doesn't have much to offer
Microsoft eats more humble pie in enterprise software business
Microsoft: selling enterprise software is a "humbling experience"


Anonymous said...

Problems with 30 - 100 users SQL I think you are confusing native Navison C-side issues with SQL - view mbs case studies Inchape = 200 navision users...

Anonymous said...

The speed degradation is similar on both native and SQL, however, on SQL it is less prone to locking problems at higher user counts.

The question about Inchape is
a) what are the roles of these presumably 200 concurrent users? It is very common with Navision to advertise high user counts, most of whom use some simplistic add-on functionality. If these 200 users take orders, ship and invoice goods and take reports at the same time, that would be nice news that it works OK
b) what is the hardware used?