Database management is one of the most important disciplines in all of IT. So if you’re reading this, you probably know quite a bit about market-leading DBMS such as Oracle. But depending on your job description, you may not know as much about some other important data management technologies. Here are five that, if they’re not already on your radar, probably should be. Data warehouse appliances and columnar DBMSdata warehouse appliances, and columnar DBMS. Both technologies are much cheaper and simpler for data warehousing than general-purpose DBMS. And if your database is sufficiently large (say, 10-30+ terabytes), managing it in Oracle may not be feasible at all.
DBMS such as Oracle and Microsoft SQL Server are optimized for OLTP, and hence for applications that have lots of random disk reads and writes. But despite huge investment, they’re not good at data warehousing. Anybody who thinks otherwise probably isn’t very familiar with two alternatives –
Mid-range OLTP DBMS, especially open source
Even if you’re not doing data warehousing, market-leading DBMS may not be the way to go. The cash cost of their licenses and maintenance may be forbidding. So may be the labor cost of managing them.
Simpler and cheaper DBMS exist that can handle ever-higher fractions of what the market leaders do. MySQL can do a lot; PostgreSQL can do more; Postgres Plus can do more yet (including running most apps that were written for Oracle or MySQL). And Progress OpenEdge has been running highly complex applications for 20 years.
Caching and memory-centric OLTP
Sometimes, the best way to do database management barely involves disk at all. At the end of the day, you almost always want to store data persistently. But DBMS that don’t routinely write to disk are useful in ever more niches. In particular, both Oracle and IBM have acquired in-memory DBMS (TimesTen and solidDB respectively) and refocused them on the middle-tier database caching market.
XML programming
The relational model has been dominant for over 20 years. But it has its weaknesses. If schemas are variable, or a high fraction of all potential data is missing, some kind of hierarchical model may be better. And that happens a lot in customer-profiling applications, because different customers give you different kinds of information about themselves.
XQuery is an intriguing way to program applications that use such models. And native XML database management systems – or native XML add-ons to relational DBMS – are emerging.
Related links:
- 4 reasons to reduce your Oracle usage
- Coverage of data warehouse appliances
- Coverage of columnar database management architectures
- Coverage of caching
- Coverage of memory-centric data management in general
- Coverage of native XML database management
