Comparing Database Management Systems (and how not to)

December 2005’s issue of Australian Technology & Business magazine (published by ZDNet Australia) compares the Oracle, IBM, Microsoft and MySQL offerings.  Somewhat surprisingly to me, I noted that Steven Turvey of RMIT IT Test Labs compared the entry level of all four products.  This result meant that the $8,000 ($206 per user + $1000 per server) Express Edition of IBM DB2, the free version of Microsoft SQL Server Express (Beta2), MySql (which is free, so hence the top-of-the-line version), and Oracle 10g Release 1 Standard Edition ($19,814 or $396 per user).

Usually I am very much a fan of RMIT Test Labs, but in this case I am not too sure about the comparison made, given the price differential.  In particular, it is interesting that MySQL won the Editor’s Choice award for the scenario set out in the review.  In particular, Microsoft SQL Server came out at the bottom of the pile, which is surprising to say the least, although perhaps to be expected given that the express edition is the deliberately-limited version of SQL Server.  A more relevant comparison, to my mind, would be the actual next step up of SQL Server – particularly since the review points noted the limitations of the express version.

At any rate I would need convincing that MySQL was the ‘winner’ given issues such as:

  • Difficulty of finding someone qualified to operate MySQL to employ to operate it.
  • The lack of available, qualified, support staff
  • The lack of commercial-grade software that is based upon MySQL such that you will inevitably need to have at least two databases that use it (e.g. your custom software has been built for MySQL, but your accounting software uses MSSQL – suddenly, you have to have two skill sets in the same application arena)
  • The fact that until this last version (Released 24 October 2005), MySQL did not have support for the standard database functionality requirements such as views (!), stored procedures (!), triggers (!), cursors (!), and a whole mess of other things – when the competitors discussed have supported these things for a Very Long Time.

I am a big supporter of open-source software, but reality does need to prevail.  The ‘new features’ of MySQL should have been there from the start – views and triggers in particular are core elements of Codd’s original outline of a relational database.  It would be a brave consultant who recommended to a client that they adopt a brand-new database management system that has had core database functionality added in only the past three months.

I think MySQL is a very good product, but I am not too convinced that it will suit all businesses equally well, and would be concerned if any businesses leapt from their commercial software to MySQL without thinking it all the way through.  A very good product, and this review is prompting me to take another look at MySQL, but I think it’s still a watching brief if the data to be managed is in any way critical to the business.

Bigger, Better, Business Intelligence

SearchCIO’s Wayne Eckerson has has just issued an article on “Five Things You Should Know about BI” (where, if you don’t know, BI stands for Business Intelligence).

In essence, he is confirming the maxim that people issues are what get in the way when it is time to play in the world of business intelligence: Politics, Lack of available tools, Culture, Costs, and Business volatility are the prime culprits.

It is interesting though to note Philip Howard’s viewpoint (from Bloor Research) over at IT-Analysis.com in his article “Do We Need Bambi?” – where Bambi is built around a hypothetical merger of “Business Activity Monitoring” and Business Intelligence.

As if we didn’t have enough acronyms in IT – now they’re breeding.

SourceForge By The Numbers

At the presentation on Tuesday night (on the commercial issues of Open Source software) I was asked if it was possible to manipulate the Sourceforge rankings. Unfortunately the website was mostly down at the time, so it wasn’t possible to answer immediately.

However, I have taken a quick look tonight and note that the Sourceforge.net rankings are able to be manipulated if one wanted to do so, as the formula is quite clear. However, as Sourceforge.net notes, those statistics are not the only way by which a project should be assessed. The ranking statistics are a good indicator of the project’s activity level rather than the quality of that activity.

The point should be made that the incentive of an project author to manipulate the rankings process is fairly low given that, in general, the potential monetary gain would be fairly minimal.

Pits, Traps and Windfalls of Open Source Software (For Business)

One of the things that I have often come across when consulting with clients is, obviously, the phenomenon of open-source software, and next week (17 May 2005) I will be presenting to the local CPA Australia IT discussion group on the topic of Pits, Traps and Windfalls of Open Source Software.

Now, I happen to think that open source software is better than the proverbial sliced bread on a picnic, but it does come with some real dangers hidden with its benefits. A real commercial issue is that, for software that is “free”, no purchase order is required and a business can find itself heavily reliant upon the open source software (and the skills of the person who knows how to use it) without any of the usual gatekeeper controls to ensure people understand what it’s all about (many businesses require a business case to purchase new software – but, no outlay means no business case means no commercial considerations are part of the decision).

And once you get out of the top five or ten open source projects in a particular software category, your ability to find someone that can actually use the software decreases markedly (which usually means that, once you find them, you’ve got to pay them quite well thanks very much). So fairly soon, and without any real red flags to indicate that it’s happening, the business can become very reliant upon the skills of one single solitary person (who may or may not be a good bloke, but is still susceptible to the all-too-common “hit by a bus” problem).

But, I use Open Office at home (fairly seamlessly for most documents) and we do sponsor open-source software such as DotNetNuke to our clients, as it’s a category killer in open source portal tools, and is based upon some standard technologies. I think it will always be interesting to run the numbers for clients and see which way they are better off. And this is exactly why I’m presenting next week on exactly this topic. So if you’re in the Brisbane area, please feel free to drop in and say “hi” by registering and perhaps discuss the finer points or two of this topic in the business context.