Practical attack on MD5(key || fixed-length-message) as MAC?

edo1

12/20/23, 9:36 PM

I wonder if there is any practical attack on MD5(key || fixed-length-message).

0 + 3

fgrieu

12/21/23, 5:28 AM

[Related question](https://crypto.stackexchange.com/q/35721/555). Brute forcing with GPUs, FPGAs, ASICs is perhaps the most practical we have, and with enough resources could be carried [above 88 bit of key](https://crypto.stackexchange.com/q/13299/555).

Maarten Bodewes

12/21/23, 9:26 PM

I don't think we can easily attack that with current tech, but the MD5 algorithm is pretty weak - so we don't know if this will be secure towards the future. Saying that the message size is fixed is fine, but beware that you could create a length extension attack if you don't verify the size **before** you or directly after you verify the authentication tag that you generate. In general you would try and not operate on the message *before* verification, so beware of implementation issues.

edo1

12/21/23, 9:52 PM

@MaartenBodewes I will not use this cipher in my program, of course. I'm just wondering if it's been hacked already.

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: Practical attack on MD5(key || fixed-length-message) as MAC?

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.