Hi, I am currently evalutating whether Hadoop might be an alternative to our current system. We are providing a web analytics solution for very large websites and run every analysis on all collected data - we do not aggregate the data. This results in very large amounts of data that are processed for each query and currently we are using an in memory database by Exasol with really a lot of RAM, so that it does not take longer than a few seconds and for more complicated queries not longer than a minute to deliever the results.
The solution however is quite expensive and given the growth of data I'd like to explore alternatives. I have read about NoSQL Datastores and about Hadoop, but I am not sure whether it is actually a choice for our web analytics solution. We are collecting data via a trackingpixel which gives data to a trackingserver which writes it to disk once the session of a visitor is done. Our current solution has a large number of tables and the queries running the data can be quite complex: How many user who came over that keyword and were from that city did actually buy the advertised product? Of these users, what other pages did they look at. Etc. Would this be a good case for Hbase, Hadoop, Map/Reduce and perhaps Mahout? Thanks for any thoughts, Benjamin _______________________________________ Benjamin Dageroth, Business Development Manager Webtrekk GmbH Boxhagener Str. 76-78, 10245 Berlin fon 030 - 755 415 - 360 fax 030 - 755 415 - 100 [email protected] http://www.webtrekk.com<http://www.webtrekk.de/> Amtsgericht Berlin, HRB 93435 B Geschäftsführer Christian Sauer _______________________________________
