Zombse

The Zombie Stack Exchanges That Just Won't Die

View the Project on GitHub anjackson/zombse

Under what circumstances would it ever make sense to use encryption during archival of digital information?

The use of encryption during archival of digital information could potentially lead to the inability to decrypt it in the future, but are there any situations wherein such complexity is warranted and appropriate in an archival process? For example, if you were archiving something under copyright and not owned by you, perhaps you would want to encrypt it until the copyright expired to justify fair use?

WilliamKF

Comments

Answer by Chris Adams

I think you have the primary reason in the question: legal or contractual requirements which obligate you to limit access and you want more protection against human error or technical lapses. If you had copyrighted materials it might make sense to encrypt the master files so a security intrusion which resulted in access to your file servers but not the keys wouldn't be a potential legal disaster. This might be particularly handy if your institution is not normally in the business of handling sensitive materials and the difficult process of retrofitting strong access controls throughout the system is too challenging.

There is one interesting technical aspect, however, which is worth reviewing: encryption requires you to treat the keys as rigorously as the actual archived material but the keys are considerably smaller which allows you to do things which would otherwise not be feasible given the size or sensitivity of the material. This could include using third-party services to transfer content and exchanging the keys over a separate channel, storing encrypted backups on a cloud provider or entering into a long-term partnership with a peer institution with the plan to release the keys later when the material leaves copyright or a donor restriction has expired. One particularly interesting aspect in some cases is that it may be possible to meet requirements for deleting content (horrible, I know, but in some cases it's necessary) by deleting the keys rather than having to prove that you've purged every copy of the data from your servers, backups, etc.

One practical implication: each item (however you defined that logically) should be encrypted with a separate, very strong key. Large data files at rest are both hard to change and easy for an attacker to brute-force so you want to pick a very strong encryption system & key and use a separate system to track the keys so you avoid the need to e.g. re-encrypt your entire HD video archive because someone wrote a password down or shared it. This is also a great spot to build in access control and logging if that's a requirement.

Comments

Answer by db48x

Encryption is usually used for keeping secrets, and if all you are concerned with is copyright (which only governs who is allowed to distribute a work), then the material is probably not secret.

I can see it being useful for a corporation that is archiving or backing up its internal communications, source code, etc. You could also use it whenever you have to store your data in the cloud, since in that case you cannot trust the provider not to get raided or shut down, even if the government is after someone else.

Just remember that encrypted files cannot be compressed or deduplicated, although you can of course compress the data before you encrypt it.

Comments