Hello, during the summer I'll have to pick a dissertation subject for my CS bachelor, and all I know is that I'm interested in databases. So if someone has experience, I would like some abstract proposals. Thank you.

My advice is to do a project and not a dissertation if possible. I.T. is a practical profession, if you want to be database admin a project (which is documentation, the app or DB) will be better than a dissertation as it shows that you can walk the walk and not just talk the talk. Dissertations are often for people who the lack the skill to do a project or lazy people as a dissertation at BSC level is only 10,000 words and projects are the documentation + the application/ DB (mine was 26,000 words + a C# app with SQL server back-end).

If you have the option i would suggest you do a project over a dissertation as you can also use that as part of your portfolio to show you have the practical skill as well as the underlying knowledge of the subject.

Thank you for your answer. Yes I'm interested in a project oriented dissertation, but I have had several "simple" projects such as e-shops or social networks (I know they can be very complex) and I would like something a little bit more academic.

you could go for the age old library system or a system for a logistics company.

Does it need to be a 'traditional' database? the noSQL movement is growing and would still be a novel project to do.

It could be anything, and noSQL is a good idea.

You could always write on Hadoop and "Big Data". That seems to be all the rage in academia these days.