New synchronization protocol for new storage backend.

[jfinkhaeuser]

Relates to #65

The general approach is the same for either implementation of the linked issue:

1. Read the content IDs to be stored from chain.
2. Read the stored content IDs from storage backend.
3. Remove content IDs that should not be stored, and use the backend's strategy to download content that should be.
4. Update storage relationships appropriately.