Previous Topic: Namespace Partitions

Next Topic: How DXmanager Uses the Topology to Configure the Backbone

Reasons to Partition a Namespace

Consider dividing a namespace into partitions if any of the following are true:

The size (memory usage) of a namespace is too big for one DSA

Partitioning the namespace amongst DSAs is the only effective way to deal with this problem.

One subtree has a much higher number of operations than other subtrees

Placing the subtree in its own partition means DSAs can be dedicated to serving just the busy subtree, and queries for other areas of the namespace are served by other DSAs. This usually improves performance

A subtree is particular to a specific business function or geographical area

This lets you maintain and control only the DSAs that serve that subtree, set up different security for those DSAs, or locate that subtree on a different set of hosts than the other areas in the namespace.