database for web analytics

what database to choose to store information about site visits, key characteristics: big amount of data, many page requests 开发者_Go百科per second, different reports for data presentation, i think to use MySql, any suggestions?

Consider letting the server log the requests and parsing them asynchronously. You don't need ACID for analytics, and you don't need to process them while talking to a client.

Most mainstream databases are good for that (including mysql, postgres, oracle etc). MySql is fine though, especially if you've used it before. Be sure look at licenses as well: MySql is GPL (the database and the connectors), Postgres is BSD, Oracle (and a few others) you need to pay for.

Most web analytics companies use some kind of distributed file system to store logs, such as HDFS, QFS... The reason is that the data is too big for the traditional database.

Analytics reports are generated via MapReduce job.

If you want to do an adhoc query, you normally use something like Hive/Pig/Sawzall.