The Passion of the Database
The Passion of the Database
In one of my recent articles I described, what I view, as a dim future for IBM DB2 in the Linux, Unix and Windows market.
I received a number of e-mails and was directed to a number of discussion threads where the article was debated.
By some I was lauded as being "an excellent writer with savvy insight and clarity of the pen." Still another described me to be "a magazine troll." Well, you cant win them all.
What struck me was the level of fervor that still exists among those in the trenches.
One responder even sent me a lengthy response claiming that my assumptions were true only if I didnt include Informix in that mix, certain that Informix would once again take its rightful place among the leaders in the relational database market.
Well everyone has an opinion, or are at least on some very strong hallucinogens.
But its all good. I used to be one of those, especially when I worked on the vendor side, where you have to drink the Kool-Aid.
In fact, it took me a while to adjust to being an industry analyst, because you have to be agnostic (or at least I believe you should, a philosophy not universally shared in the analyst business I have found).
You move from being a fan and cheerleader to being an observer without passion or prejudice.
Its this passion that leads some organizations to make potentially bad decisions. Not bad from the standpoint that total failure occurs, but more along the lines that equally good yet more cost-effective choices might have been missed.
Now, lets be honest, buying decisions (in IT, anyway) have more to do with risk avoidance (which translates to market share) than they do with the relative strength of any product.
When you have passion for a product, you lose sight of this facta lesson I learned many years ago when speaking to a customer that had chosen a competitors performance monitor over my companys.
"How could you make such a choice?" I said incredulously. "Our monitor tracks over 130 metrics, and theirs has only 60!" The answer hit me like a truck, an epiphany really, "Well theirs had the 60 we cared about."
So "best" is a highly subjective term that means that the "best" product will always be a loser in some peoples opinion.
As for DB2 for Linux, Unix and Windows By the way, the name alone pisses me off.
I mean, its so long, yet I have to be specific, lest I confuse people with the other database platforms that IBM has also named DB2 but are clearly different.
In any case, I digress. DB2 for Linux, Unix and Windows, readers should not interpret from my words that IBM is pulling out of the database market or that I am attempting to influence the market because I own stock in Microsoft or Oracle.
In the spirit of full disclosure, its possible I do, but who reads the prospectus mutual fund companies send out anyway? Once again, I digress.
The truth is that DB2 for Linux, Unix, and Windows (DB2 LUW) has no market momentum, which translates into: Its riskier than other database options.
The truth is also that while DB2 LUW, Oracle, and Microsoft SQL Server may all be comparable options for most things, it does not make databases a commodity.
For one thing, switching costs can be very high. So high switching costs and low market momentum tend to be against DB2 LUW.
Indeed, in one survey conducted this past spring, almost twice as many users identified MySQL versus DB2 LUW when asked which database was the fastest growing in their environment in terms of applications supported.
Next Page: Fluid future.
I dont believe that this prediction is inevitable, however. The future is a fluid concept (which makes it great for those of us that predict it), so I could be wrong if IBM were to begin acting as if it cared about its database business as a growth engine.
I suggest IBM play to its strengths. It is much larger and significantly more revenue-diverse than Oracle.
IBM should begin to use this to its advantage the way Microsoft does so well. Shake things up by attacking Oracle where it is most vulnerable, licensing and pricing.
This is a land grab for market share, so IBM needs to do what is necessary to get DB2 LUW installed.
I suggest they do so by eliminating license fees for DB2 LUW for servers up to four processors. Im talking about enterprise edition here, folks.
This eliminates the confusion I spoke of in a previous column about database licensing complexity.
This enables the market to easily download and freely install an unquestioned, top-of-the-line RDBMS.
The four-processor limit still covers at least 80 percent of the database workloads running today. Im not talking about IBM open-sourcing the code; this is simply a licensing and pricing decision.
Where does IBM make its money? Support services.
Unlike MySQL, the vast majority of global 2000 companies would instantly become interested in deploying DB2 LUW for future applications or when significant application upgrades are needed.
From my discussion with these types of companies, none of them are interested in going it alone without a support contract.
Margins on support contracts are very strong, and since IBM is not saddled with the install base of an Oracle, they could grow revenues significantly.
Once installed, IBM has an opportunity to keep selling into its install base with options such as data partitioning.
There is also the revenue opportunity from customers who may need to run DB2 LUW eventually on servers larger than four processors (not cores).
More installs means more people picking up DB2 skills, which only adds to the momentum.
How would Oracle respond? Well, with more than 70 percent of its revenue tied to its database software sales, they better get real good at selling ERP applications.
Microsoft could still hang, but such a move would also deal a blow to open-source database growth, or at the very least slow down its momentum and shift all the buzz to DB2 LUW.
In any case, that would be my suggestion, but who listens to an agnostic, anyway?
Charles Garry is an independent industry analyst based in Simsbury, Conn. He is a former vice president with META Groups Technology Research Services. He can be reached at firstname.lastname@example.org.
Check out eWEEK.coms for the latest database news, reviews and analysis.