SSIS: Why I’m Standardizing on ADO.NET Connection Managers

When connecting to a database with SQL Server Integration Services, you will most likely use ADO.NET, OLE DB, or ODBC.  Until I find reason not to, I’ll be standardizing on ADO.NET connection managers as much as possible.  Even if the other database, such as mySQL, has only an ODBC connection, I’ll use ADO.NET instead and wrap that around mySQL’s ODBC connector.  My reason for this is simple:  I use a lot of script tasks and script components in my SSIS packages.  It is far easier to work with connection managers in script tasks/components when those connection managers are ADO.NET.  Consider the following:

  • OLE DB and ODBC, from what I can tell, do not enlist the current transaction.  What this means is that if you, for example, have an Execute SQL Task that starts a database transaction (assuming the database connected to supports it), the script task/component will not handle it’s steps in that transaction.  The most common occurrence is your package crashes.
  • OLE DB and ODBC when used in the script task/component do not support the “Retain Same Connection” setting. 
  • The code to acquire an OLE DB connection is not as elegant as the ADO.NET connection.  This is the number one reason why I am standardizing on ADO.NET.

I’ll have more to say on the third part later on – which is the most important to me.  Now that Microsoft has finally fixed the Script Component such that you can debug into it starting with SQL Server 2012, the Script Component is now very useful.

JamesNT

Why Windows Server 2003 Will Be Around A While Longer

Hello, Everyone.

It’s been just over a year since my last post on my blog.  Things got really interesting last year both on the professional and personal fronts.  As time moves on, I’ll discuss many of those things here.  To start off 2015 with blog posts, I thought I would cover something I’m seeing a lot of that we saw last year:  The end of support for a major Microsoft product.  Last year it was Windows XP.  This year, it’s Windows Server 2003.

There have been lots of articles on the web recently about the end-of-life of Windows Server 2003.  Like its little brother, Windows XP, Windows Server 2003 is to have support completely and permanently removed by Microsoft unless you are willing to pay Microsoft some additional money to extend support.  For more information on the Windows Server 2003 lifecycle, as well as the lifecycle of other MS products, visit the Microsoft Product Support Lifecycle page.  The problem I have with most of these articles is they generally fall into two categories:

  • People just don’t want to move; therefore, are being irresponsible.
  • It’s the economy, stupid!

But there is a third reason that I have seen and this third reason is more prevalent that any of the previous two mentioned above, in my opinion, as to why so many are delaying their migration off of Windows Server 2003:  Because there are still lots of applications out there that still will not run on more modern versions of Windows.  Especially applications written in VB6.

Windows Server 2003 is the last server operating system that will comfortably run VB6 applications without much fuss.  Once you hit Windows Server 2008 it’s pretty much game over for those applications.  While Microsoft does offer limited support for VB6 all the way through Windows 8.1, the problem with every VB6 app I have seen is that they break so many development rules that getting them to run on any version of Windows past Windows Server 2003 is practically impossible.  More modern versions of Windows have higher security standards and so forth that just won’t allow an errant application to do whatever it wants.  I realize that Microsoft has tools available to help with these things, but the point stands that getting many older applications to run on newer versions of Windows is painful, expensive, and the application may yet run less stable than before.

Some of you may wonder aloud why a company would still be on an application written in VB6 or, at the very least, old enough to not run on later versions of Windows.  Because moving line-of-business applications is HARD.  I’ve done this before, am in the process of doing it now, and I can tell you it is HARD.  Please consider the following:

  • Many line-of-business applications today are much more expensive than they were years ago.  A great example is the healthcare arena.  Practice Management and Electronic Medical Record systems that cost ~$5,000 10 years ago are well over ~$20,000 today.  That’s a big jump and most certainly hard to swallow.  One can’t help but wonder how many other applications cost more today than they did yester-year.
    • Along this line goes lack of expertise.  Just because you are on version 1.x of a software and need to go to 12.x doesn’t mean it’s easy.  You may need to phase your upgrades across virtual machines and so forth.  For example, what if the line-of-business software requires a different version of a database when going from Server 2003 to Server 2012 and you have to convert data?  What if the line-of-business application changed databases entirely (e.g. going from Advantage database to SQL Express)?  Who is going to handle all that?  Consultants with that kind of skill can cost a lot of money.  Congratulations, you just doubled the cost of your upgrade.
  • Moving to a new version of a line-of-business application may involve massive re-training.  With new features and changes to the user interface, staff may have to learn their way around all over again.  A good example is the big change from Office 2003 to Office 2007 with the advent of the Ribbon.  Personally, I love the Ribbon in MS Office, but nonetheless it required lots of retraining.
  • Your line-of-business application vendor may not exist anymore.  A lot of companies have gone bankrupt thanks to the housing crash of 2008.  Your company may have to switch line-0f-business application vendors entirely and that is a whole new ball of wax.

I could go on but you should see my point by now.  Even companies with very strategic plans on handling IT and software deployments are finding themselves in a crunch with Windows Server 2003.  Microsoft may find itself selling some extended contracts for many.

It may be time to be a bit more forgiving towards those that are still on Windows Server 2003 for a bit longer.  And, if you are a consultant familiar with older technologies and their newer counterparts, it may be time to start a new advertising campaign.

JamesNT