Comments on: What leading DBMS vendors don’t want you to realize

By: DBMS2 — DataBase Management System Services » Blog Archive » Naming the DBMS disruptors

Fri, 25 Apr 2008 04:04:43 +0000

[…] Edit: This post has largely been superseded by this more recent one defining mid-range relational DBMS. […]

By: Curt Monash

Curt Monash — Mon, 28 Jan 2008 22:50:52 +0000

Lux,

Sorry, but I wouldn’t know. Unless it’s changed names over the years, this is the first I’ve ever heard of 4D.

CAM

By: lux

lux — Mon, 28 Jan 2008 22:30:51 +0000

Curt,

Would you categorize 4D as a mid-range DB? If not, why not?

By: Log Buffer #81: a Carnival of the Vanities for DBAs

Log Buffer #81: a Carnival of the Vanities for DBAs — Fri, 25 Jan 2008 17:52:30 +0000

[…] Monash of the DBMS2 blog tells what leading DBMS vendors don’t want you to realize: “For most applications at any enterprise – and for all applications at most enterprises […]

By: Curt Monash

Curt Monash — Thu, 24 Jan 2008 19:40:56 +0000

Bob,

So are you arguing that everybody is delusional about these new features? Or are they indeed useful for somebody?

Best,

CAM

By: Curt Monash

Curt Monash — Thu, 24 Jan 2008 19:39:55 +0000

Dan,

That’s the classic Clayton Christenson “disruption” narrative, as applied to software. I summarized it in six bullet points in http://www.dbms2.com/2007/02/27/oltp-database-management-system-disruption/

Best,

CAM

By: Daniel Weinreb

Daniel Weinreb — Thu, 24 Jan 2008 13:41:32 +0000

Even Oracle is starting to realize that many applications don’t need the full power and cost of Oracle Database, as I think we can see from their acquisitions of SleepyCat and TimesTen.

What’s happening to Oracle in databases is the same thing that has been happening to BEA Systems (WebLogic) in J2EE application servers. New entrants, many of them open software, are getting better and better, and doing more and more of the things that the high-end products can do. The high-end products are forced into a smaller segment of the market. It’ll be a long time before BEA is gone — lots of companies currently depend on BEA’s software and extremely reluctant to switch — but growth in WebLogic Server itself is clearly a big problem.

I feel like the same thing happened to InterLeaf, many years back, as FrameMaker started to push up the ladder, and then Microsoft Word and Publisher started pushing up FrameMaker’s ladder.

Getting forced up into the high end, where you have to provide lots of fancy and esoteric (read: hard to develop and hard to Q/A) features and high performance and such, but you can only sell it to many fewer customers, is not a nice place for a software vendor to be in, especially if they can’t raise prices easily.

By: Bob Zurek

Bob Zurek — Thu, 24 Jan 2008 13:07:27 +0000

Let’s keep in mind that most of the big dbms players have to deal with very complex and time consuming internal processes as they pursue a new release schedule. These schedules are frequently extended due to these complexities and also get extended because of what I call “feature wars” that go on inside the big software company, especially those companies that also sell applications that best run on their own database. Conflict is created because the application teams want certain features (remember the row locking versus page locking wars of the past?) while the product managers, competitive intelligence, customers and even sales/marketing teams want their “favorite features” to respond to innovation requests, customer requests and competitive pressures. It’s a tough cycle that many software companies experience as they become big, bloated and very complex. This on top of the never ending push to get a release out the door for revenue recognition and upgrade revenue desires. A much different cycle takes place on open source projects and that can be a huge benefit to someone who takes advantage of an industrial strength proven OLTP database like EnterpriseDB and PostgreSQL. How long did it take to get some of the big database vendors releases out the door. How many of the 100’s of new features in the big database vendors product are really taken advantage of for what might be a complex upgrade process? Oh, and if you are going to go thru an upgrade or thru the process of creating a new app that requires a database, why wouldn’t you try an open source database like EnterpriseDB or PostgreSQL?

By: Curt Monash

Curt Monash — Wed, 23 Jan 2008 03:38:17 +0000

David,

You’re making me feel old. The partitioning wars — static vs. dynamic — were huge in the first half of the 1990s. In those pre-web days, I consumed a LOT of ink writing about them.

CAM

By: David Holoboff

David Holoboff — Tue, 22 Jan 2008 22:05:08 +0000

Think of the reasons behind high end performance in Oracle DBMS, and check those against the mid-range DBMS’s today and you will be able to determine effectively what you need.

Take partitioning for example: Without partitioning, Oracle would be crawling with a few of the tables we had. However, the world was significantly altered with partitioning in place. Same will go for other DBMS’s when they offer partioning (along with simplified maintenance).

When mid-range DBMS’s continue to offer more of these reasons behind high end performance, then the need for the high-roller solution will continue to diminish.