Transparent Database Encryption in SQL Server: A Planning Guide

  • Comments posted to this topic are about the item Transparent Database Encryption in SQL Server: A Planning Guide

  • Very good effort Ron. You have handled a tough topic with lot of care. Thanks.

    M&M

  • "Symmetric algorithms require a single key for both encryption and decryption which allows for high-performance;"

    I would say both facts (using single key for encrypting and decrypting data and algorithm speed) has nothing in common.

    Symmetric algorithms use the same key for encrypting and decrypting. Period.

    Widely used symmetric encryption algorithms are faster than asymmetric ones. Period.

  • It's quite frightening. I don't like to go full weight into criticism of contributions, but here I feel I have to. The author obviously knows far more about how encryption has been incoprporated into SQL Server than I do. Despite that I think he may not get people to think about the things they need to think about, and I don't think he will help people understand much about encryption itself as opposed to how it fits into SQL server, and may instead mislead.

    I look at diagrams 1 and 2: one of the apparent differences between symmetric and asymmetric systems is that symmetric systems, unlike asymmetric systems, require an algorithm, rather than just a key, to do decription. That's pure rubbish. And then it goes downhill (I was quite surprised to discover that that was possible) from there!

    Let us compare a 128 bit asymmetric key using our best asymmetric algorithms with a 56 bit key pair using our best symmetric algorithms; which is more secure? Well, if you believe this article, the asymmetric key pair is - and if you believe that you are wrong by a few (if "a few" is allowed to include "a lot of") orders of magnitude. Why is this nonsense statement about the relative security of the two encryption types included without any reference to key lengths? I can factor a 128 bit number for nothing comapred to the cost of breaking a 56 bit DES key, and the latter is generally regarded as pretty insecure.

    There's anice descrition of symmetric vs asymmetric key systems. It would be hard, however, to discovet from this article that SQL server uses symmetric key encryption to encrypt data, and asymmetric key encryption to protect the symmetric keys. That strikes me as pretty fundamental - but it's either not there or so well buried that I didn't spot it.

    The most important issue of all in cryptography is not addressed at all - and this is supposed to be a planning guide, so that's thoroughly appalling. That most important issue is who knows the keys and who knows the passports to get access to the keys on the system. If you have an encrypted database without a thoroughly worked out policy covering that issue you may as well abandon encryption - you are just wasting your time.

    edit: fix grammar. Cha mhór leam a bheurla Shasunnach! (English is a pain)

    Tom

Viewing 4 posts - 1 through 3 (of 3 total)

You must be logged in to reply to this topic. Login to reply