Improve / investigate history data storage format

[kdeme]

Follow-up issue of https://github.com/status-im/nimbus-eth1/issues/1083

We have a json & db storage format for history data (block headers, bodies and receipts).

We should look into improving the current structure: e.g. with the current structure one can not iterate over a range of blocks from a file, however one can easily access one specific block based on hash.

And perhaps also explore other formats. E.g. currently data will be written either to json or to a "temporary" database, for then to be used in fluffy. Something like e2store could be used that would potentially fulfill both scenarios