Presently the source code that is analyzed and stored in Sonar's database is stored uncompressed and in the clear. As databases grow this requires an enormous amount of disk space to house this data. I propose that this data is compressed prior to storage and expanded when returned to the UI. Whilst there will be a processing overhead in doing so, the benefit is that databases do not become over-bloated. Obviously not all the databases that Sonar supports provide CLOB/BLOB encryption, so I would suggest using a database agnostic mechanism such as the Zip classes in Java.
This will also enable in the future the ability to encrypt the source so that it is no longer stored in the clear.