IBM Cognos Dynamic Cubes

28 IBM Cognos Dynamic Cubes

The result set and expression caches, in turn, reduce the amount of processing that occurs

within the MDX engine within the DQM server, helping to reduce spikes in resource usage.

Cognos Dynamic Cubes also makes use of available processing cores to do costly operations

in parallel, such as loading hierarchy tables, or processing query result sets.

Cognos Dynamic Cubes uses five caches, each with a separate purpose. The following

sections describe these cache types.

2.3.1 Result set cache

The result set cache exists at a layer directly above the MDX engine with the DQM server.

The result set of each MDX query executed by the engine is stored within the on-disk result

set cache. Each MDX query that is planned for execution is first searched for in the result set

cache. So, if the query was previously executed, and the combination of security views of the

user who executed the query matches that of the current user, then the result set of the query

is obtained from the result set cache. Thus, the query is not executed other than the first time.

Figure 2-23 shows the result set cache.

Figure 2-23 The result set cache

The cache is aware of queries that are currently executing, so it is not possible for a query to

be executed twice. If necessary, a second request will wait for the first to complete execution.

Also, the result set cache is flushed when the data cache is refreshed to ensure the data and

result set caches are synchronized.

Although stored on disk, the result set cache is capable of providing extremely fast report

execution because, in many cases, the output of a report is small compared to the amount of

data that is processed to produce the final result set.

The result set cache, besides sharing report execution results between users, also speeds

drill up operations, because the drill up involves the re-execution of a query.