Okay, first thing when designing a database is to know what it is you are looking to do. You are looking to create a music database so we need to understand all the different entities used in order to model our database. Entities are the objects that we simulate, they can be physical items such as musical bands and abstract items such as genres. Each entity you can identify is a way of identifying each table necessary by default for your database. Entities are the basic naming part of a group and then, under entities, we have their properties which form our fields i.e. a band's name, year formed, year split etc. are all properties.
So, when we think about it the following are our entities:
- Artist
- Song
- Genre (Rock, Pop, Jazz, etc.)
- Format (Cassette, LP, CD, etc.)
- Record Company
So, we have 5 tables to begin with. Now, when we look at some of these we realise that more than one artist can sing a song and one song can be song by many artists. Therefore we have a many-to-many relationship. This can only mean we need a new table to simulate this. So, we add a new table for
artists to songs. Record companies, likewise, can form a many-to-many relationship with artists. We need a table for this join. We keep doing this until we are satisfied all our relationships are correct.
For our record company we'll need their name, address, phone number, email, etc. We don't need the bands that come under their umbrella - that's done in our joining table.
I don't have time to go much further into this in detail but I hope that can you a start in where to look. The keyword in all of this is
normalization - search the web; there's loads of tutorials on this.
Unsolved 
I recently did that assignment 
Offline 
