No one would argue that DB2 is one of the leading (R)DBMS products in the market. As any other powerful software product it has complex architecture. Here I’d like to introduce some basic terminology of it in series of posts because sometimes this terminology can become somewhat vague.
Lets kick off by explaining some general view of DB2. Each IBM DB2 installation has following basic levels of architecture:
- Instance can be understood as a completely independent environment with it’s own security configuration, resource allocation and contains databases and partitions isolated from all other instances. Each instance has its own system processes which manage data. Instance may contain several databases.
- Database is a most familiar term. It’s a logical unit which holds your data. It has complicated structure which we will explain later on. Database can reside in one or more partitions.
- Partitions (or nodes) is a way of creating DB2 database cluster for the sake of higher performance. You can split your database onto several servers where each of them will have their own chunk of data. Since more servers means more CPU cores, memory and disk I/O it’s a natural way of scaling DB2. Particular storage resource is called Container.
- Database is then mapped onto several Tablespaces. Tablespace allows you to manage how database tables are held on your storage resources. For example you can hold your data in files as well as on raw hard drives (which is faster). In addition it’s possible to put frequently used data on SAS hard drives and rarely used on SATA by means of creating two different tablespaces. On top of that you can set up a different page size for each of your tablespaces.
Here is the basic idea of what DB2 represents from the view of systems administrator. To better understand a folding of different levels please refer to this well-known among dbtwoers picture: link.
Tags: container, database, DB, DBMS, IBM, instance, node, partition, RDBMS, tablespace
Niktips's Blog 
Leave a Reply