Talk:Data Encryption Standard

Key size

The article currently has "In the submission of proposals to the U.S. government, IBM proposed a 64-bit key". Can we document that? Other references I've seen, including one I've just added to the article and Applied Cryptography p 267, all give Lucifer key size as 128 bits. This does not necessarily mean the quoted assertion is wrong, but it does raise questions. Sandy Harris 02:33, 3 November 2008 (UTC)

Well, I wasn't involved in all the negotiations, but, at the time, I was the Library of Congress representative to the Federal Telecommunications Standards Committee. The NIST and NSA representatives all said IBM wanted 64, but the government recommendations were 56. It was also explained that after negotiations with NSA, IBM had accepted that the details of the S-box design should be classified.

I don't know if it's online anywhere, but the Senate Intelligence Committee did an extensive review of the issue, including establishing its own cleared expert panel -- I remember Dorothy Denning was one of the members. It was published by the Senate, and it addressed the specific point of NSA wanting 56 when IBM had recommended 64. If we can find someone with a copy, or if it's been scanned, that should pin it down. Howard C. Berkowitz 04:08, 3 November 2008 (UTC)

I cannot find the Senate report. Checking other references, the Menezies et al Handbook says 128, as do many others [1], [2] [3] [4]. The only exception I find is Koops' thesis [5] which says 112. Sandy Harris 09:26, 3 November 2008 (UTC)