Use CBOR for database persistance?

Hi there,

I love this crate - I've done some experimentation between using JSON vs using CBOR as the database persistence format, and I've decided to use CBOR within my own application for several reasons:

1. Database size: CBOR is about 1/5th the size of JSON when storing memvdb databases
2. Loading / Saving speed: CBOR is about twice as fast for saving / loading database
3. CBOR is an IETF standard just like JSON ([RFC 8949](https://www.rfc-editor.org/rfc/rfc8949.html)).
4. CBOR is considered the "successor" to JSON

I'm using the `serde_cbor` crate.

It might be a good idea to move to CBOR as the default serialization format for the above reasons - but ultimately, since memvdb generically supports serde, it could also be left up to users.

More information about CBOR can be found here: https://cbor.io/

No files, tests, or entry points are identified. First review how memvdb currently serializes and persists databases, then compare the implications of making CBOR via serde_cbor the default versus leaving serialization to users. Done requires a decided scope and an agreed persistence-format change.

Written by the indexing model from the issue text.