This might be a dumb question since I have not yet fully understood how Cognos BI works. Also, I tried posting this on Stack Exchange, in vain (error with putting tags).
My question is - Can Apache Hadoop be used to make Cognos BI work faster ? Or is Cognos doing the same thing which Hadoop does (MapReduce functionality) ?
The place where I have started working, uses the Cognos BI suite on top of Sybase IQ (the content store). Apache Tomcat is the web server. What happens sometimes is that Cognos takes a lot of time(almost dies) to generate reports if the data set is large.
So can Apache Hadoop help Cognos perform better by fitting somewhere between Cognos and Sybase ? Or is report optimization the only way out in this case ?
Thanks Guys.
Hadoop as a platform is not aimed for ad-hoc queries or analytic reports.
Cognos is an IBM product. It can only query it's own distribution of Hadoop, which is called big insights:
InfoSphere BigInsights
Over BigInsight Cognos issues queries using Hive, which eventually translate to MapReduce.
You say you are using Sybase IQ (this is not the content store, this is the reporting DB your queries are running on).
Although I don't know much about Sybase IQ, I am working heavily with Vertica, which is also columnar DB.
In order to get good performance, you have to tune anything possible: