5 ms·
From my very not-a-crypto-person perspective, I had thought compression was fine as long as you don't pad it to the expected block length before encrypting? so
by coroutines 10y ago
From my very not-a-crypto-person perspective, I had thought compression was fine as long as you don't pad it to the expected block length before encrypting?
something something padding oracle
- gcommer 10y agoThe paper linked in the article: "Phonotactic Reconstruction of Encrypted VoIP Conversations"[1] specifically works in a situation without padding because the compression scheme affects the length of packets in a semantic way. [1] http://www.cs.unc.edu/~fabian/papers/foniks-oak11.pdf http://www.cs.unc.edu/~fabian/papers/foniks-oak11.pdf