Tom Yagers Reality Check

Infoworlds Tom Yager has given in his Ahead of the Curve column in InfoWorlds January 16th 2006 issue his tacit blessing to “it runs best in Windows”. Tom calls it “better together” and cites the triumvirate of Windows Server2003, Visual Studio 2005, and SQL Server 2005 as candidates for running better together. He does acknowledge some of the risks of “Better together” citing “strong armed upgrades and whole catalog purchases…Sounds a lot like lock yourself in youll love it. And yet slowly Microsoft is making togetherness valuable for customers.”

Tom goes on to cite that Windows Server + SQL Server 2005 + Visual Studio 2005 has to meet the needs of Microsofts toughest customer – Microsofts own IT department. And that departments needs can act as a good proxy for all large IT departments. And Microsoft IT has finally seen the light after 30 years in business and is finally becoming “reality based” . Microsoft IT is finally eschewing BSD, AS400, Unix and other intemperate IT choices and is going “runs best in Windows.” And this movement has legs 1 Microsoft Way wide so it should for all IT shops.

A lot of Toms assessment is based on the quality of the marks given in the review in Infoworld for SQL Server 2005 by Sean McCowan. That review gave the following marks :
Manageability -> 9 despite the fact that MS does not match IBM and Oracle in self repair, auto-tuning capabilities and is just now starting to match the comprehensive competence of DB2 GUI-based Control Center. A 7-8 rating is closer to “reality based.”
Performance -> 9 despite no benchmark testing and SQLServer lagging DB2 in TPC by factor of 3 and Oracle by 35% despite Microsoft using 64bit Windows+SQL Server for both TPC and TPH versus 32bit for Oracle and DB2. A 6-7 rating is “reality based.”
Availability -> 9 this is hard to corraborate for such a new product. J.D.Power and Associates has said it will be publishing results by mid year for enterprise IT software. Cant wait to see the database assessment procedures and numbers.
Scalability -> 9 again despite the fact that some of the key clustering and data mirroring features have yet to be delivered. Shame, shame. A provisional 7 is much closer to “reality based”.
Value -> 10 is hard to argue with but there is a real moral and pragmatic business issue to be dealt with here.
Microsoft is giving away a free and very comprehensive BI Stack (OLAP, Data Mining, ETL, Alerters and Event Handling, Reporting Design and Delivery)with each copy of SQL Server sold. This is an exact repeat of the predatory pricing that Microsoft unleashed on Netscape when it gave away “free for perpetuity” the IE browser and IIS Web Server and then had to pay AOL/Netscape $1.2B as a result of the Anti-trust trial and the predatory pricing ruling against Microsoft. Not a peep about this in either Tom or Seans assessments despite the fact that IE has seen no functional updates for 6 years and counting while perpetually free IIS is now part of the Windows XP Pro bundle helping to raise its price by $140-200 over Windows XP Home Edition. Also perpetually free IIS latest edition is nowhere to be found on Microsofts website as a free download. Given this history I am surprised that neither Tom nor Sean deemed it necessary to issue a word of caution to customers or explore with Microsoft product managers how long these bargain basement prices will be maintained.

Finally, the other major gap in both Tom and Seans assessment of SQL Server 2005 and and the “better together” trio is little mention of interoperability. That includes assessment of SQL Servers adherence to ANSI SQL 2003, XML Standards, and broader Web W3C so called 2.0 standards. The new BI Studio for example supports only Excel, HTML, PDF outputs; not the Java, Flash, CSV, XML by DTD offered by many competing vendors. In general, assesment of SQL Server 2005 and its various components against standards and interoperability criteria was remiss.

Now the reality based operation at Microsoft IT is that it is finally becoming a completely Windows shop. But that has consequences because of 3 things:
1)If Microsoft makes any big acquisition (still think SAP) there is a high likliehood that there will be major IT conversion costs because Microsoft is least interoperable of major IT vendors. Lots of servers conversions, lots of service and application server changes because Microsoft makes minimal effort to support other Application servers be they CORBA, DCE, J2EE, or EDI . Now there is also the possibility of lots of database changes as well as Microsoft database+development+server software becomes so interdependent (it is called tight coupling in impolite nerd circles);
2)In talking to suppliers, customers, goverment and partners, Microsoft will have less options to offer for exchange of information. Microsoft will be imposing its methods and processes on others – not a big change, just another burden of doing business with Redmond. Others in the marketplace cannot afford such arrogance and hubris;
3)Getting the right info in the hands of users by breaking down islands of information and application silos through data integration and interoperability are the number 1,2,3 goals for organizations. By ignoring ever wider swaths of dejure and defacto standards, Microsoft is trying to tilt the playing field towards the notion that everything runs best inWindows and interoperability is not important. But critically -Redmond can ignore and set defacto standards as it deems necessary in any market where it has 80-90% market share. Hence Open Office XML over OASIS or anything else. In the Web development this attitude imposes the 20-35% greater work for IE tax. And in the security shops this is the monthly second Tuesday scramble the cost for which I am sure all those Microsoft sponsored “Get the Flaps on Linux” campaigns have fully costed out (if you are looking for such data its with your tooth, under the pillow).

Now Infoworlds review and assessment of SQL Server is not off base. SQL Server 2005 is greatly improved over SQL Server 2000. But so are Oracle, DB2 and Sybase among others. This assessment seemed targetted more towards Microsoft shops which want to know how much gain for how much pain there is in moving from SQL Server 2000 to 2005. And Sean is mostly upfront warning that there are significant chunks of enterprise software to wait for. There are also significant changeover and learning costs which Sean, thankfully, does not blink at in DTS, management, BI Studio, etc. But this is still like the improvement in MacBook which we assess here. Sure the MacBook is a big improvement over existing Apple Powerbooks; but relative to PC Notebooks of comparable features, MacBook leaves a lot to be desired in the pricing arena. Ditto for SQL Serrver 2005 and its Windows Server 2003 and Visual Studio 2005 partners – there are a number of availability, security, scalability/performance details yet to be shown which simply do not warrant 9/10 ratings. But even more critical there are major pricing, performance and interoperability issues as well.

So it is important to note that Tom and Sean do not mention any one of these latter notions. Hence I have been arguing that in matters like this review and assessment, the trade press is acting as accomplices to the Microsoft line – turning a blind eye on major issues or giving them kid glove treatment. Such appears to be the case here.

(c)JBSurveyer 2006