Talk about Cognos' thinking of handling big data, only for the version below 10.2. 1, and don't introduce the distributed data warehouses such as hadloop introduced in 10.2. 1. We mainly start from a general medium-sized project and use what ideas to optimize our query.

We mainly think about the processing of big data from three ideas.

First, the database level

Now the mainstream Cognos project, the main development mode is dmr report modeling based on rolap. Therefore, database optimization is very important. Optimize our database mainly through the following aspects:

(1) Index, establish and maintain key reduction fields, such as dimension id and dimension level id.

(2) According to the data size, partition optimization is carried out according to time.

(3) Use of cache table MQT

(4) Tablespace and buffer pool settings.

(5) database performance optimization

Second, Cognos server optimization.

Cognos optimization includes configuration file optimization, cluster construction, service and log opening, etc. The installation and configuration optimization based on Cognos software mainly includes the following aspects:

2. 1 apache configuration optimization

Optimal configuration of timeout/maxkeepaliverequests/keepalivetimeout.

2.2Cognos comes with tomcat configuration tuning

(1)TOMCAT configuration file CrN _ root \ Tomcat. \ conf \ server.xml can be modified. Its parameters are concentrated on the line:

You can modify maxprocessors/acceptcount connection timeout.

(2) File path: CrN _ root \ Tomcat. \ conf \ web.xml。

You can modify the session timeout.