Edit: This post has largely been superseded by this more recent one defining mid-range relational DBMS.
I find myself defining a new product category – midrange OLTP/multipurpose DBMS. (Or just midrange DBMS for brevity.) Nothing earthshaking here; I’m simply referring to those products that:
- Match the OLTP DBMS state of the art of some point in the 1990s, and hence meet most application needs today.
- Are significantly more affordable (in license, maintenance, and/or administrative costs) than the top-end products from Oracle, Microsoft, and IBM.
Basically, these come in three groups.
- “Open source” – these are the first ones most people think of. MySQL leads the way – if 5.0 works as advertised. EnterpriseDB has a very interesting offering. PostgreSQL has some fans away from EnterpriseDB. And then there are Ingres, Firebird, and others, at various levels of open-source purity.
- “Standard editions” – Informix SE may have been the first DBMS to use the “Standard Edition” term, but Oracle, IBM, and Microsoft all have free or low-cost versions of their flagship DBMS. These are basically crippled versions of the lead products, conceived as entry-level products to fend off price competition.
- VAR-oriented – Progress OpenEdge and InterSystems Cache’ are the two biggies in this category. (Informix SE also used to be pretty big.) These are products that are great to leave running with a stable packaged application suite in a small enterprise, or with a single custom app in a large department. I’m visiting Progress on Monday and may have more to say after that.
Even that’s a far from complete list. For example, the Sybase iAnywhere core product used to be a great laptop/desktop/workgroup RDBMS, and Sybase makes periodic noises about recommitting to that market.
So should you care? Yes. Overall, there’s a swelling, rather classical, disruption story. And if you’re an enterprise looking to consolidate your DBMS suppliers, you would do well to use one of these midrange products as well as (if you even need it) one of the top-end OLTP RDBMS.
As for what these midrange DBMS are or can be used for – well, that’s a long list. But some of the broad categories are:
• Most web-facing apps.
• Most departmental apps.
• All the processing at a typical medium or small enterprise.
Of course, the details depend on the match between your application needs and particular products. As just one example — if you need non-tabular datatype support, such as XML or geospatial, over half the possibilities can be immediately thrown out.