IBM Cognos Dynamic Cubes

Chapter 2. IBM Cognos Dynamic Cubes architecture 27

Both metadata and data requests apply any security defined in the model before returning

member metadata and before retrieving measure values.

Data requests, through the query strategy of a cube, make use aggregate and data caches of

a cube.

Figure 2-22 shows the command dispatch within the DQM Server.

Figure 2-22 Command dispatch within the DQM Server

A single DQM server may host one or more dynamic cubes at one time. If a DMQ server

supports a virtual cube, all of the cubes upon which the cube is based must also reside on the

same server. Because a dispatcher can be assigned to only a single server group, all cubes

that reside within a DQM server are implicitly part of the same server group, and all packages

that reference any of the cubes on the server must all route their requests to this single server

group.

2.3 Cognos Dynamic Cubes caching

The basis for the performance of Cognos Dynamic Cubes is its various in-memory caches

and its use of database summary tables. The DQM server supports 64-bit machine

architectures, and, as such, provides Cognos Dynamic Cubes access to large amounts of

memory.

The aggregate and data caches in Cognos Dynamic Cubes not only provide fast query

response, but also help to alleviate the processing load on the underlying relational database,

because Cognos Dynamic Cubes does not have to re-retrieve data that is held in its various

caches.