Archiving `job` and `results` objects

[quantumjim]

What should we add?

As mentioned in #6978, it should be possible to serialize job and results objects. The reasons being:

Convenience

An internet connection cannot always be taken for granted. By allowing users to make a local copy of their job (and associated results) they can work on it more conveniently. It is easy to save certain things individually. But when one is stuck without an internet connection, one invariably needs an obscure detail within the job object that they did not save.

Sharing with collaborators

job objects are associated with a given account, but projects are often worked on by multiple people. It is more convenient for all to be able to access the data on the same terms (for similar reasons as above) which would be allowed via the serialization and sharing of these objects.

Open Data

Funding for research now often requires data to have long-term, reliable storage and made openly available. Preserving a full record of the job means preserving the job object, the backend information from the time of execution, and the results objects. Making this easy and reliable will help researchers satisfy their open data requirements.

I already made my own attempt at this, to satisfy my own open data requirements https://github.com/NCCR-SPIN/archiver4qiskit but updates to Qiskit invariably cause mean that the pickled objects no longer work correctly. A reliable solution that is actually part of Qiskit would be the ideal solution.