US kicks off secure hash competition
Gentlemen, choose your algorithms
Schneier and seven other cryptographers from industry and academia have submitted a proposal for their "Skein" family of hash functions. Cryptographer Ron Rivest - known famously as the 'R' in RSA encryption - and a team of more than a dozen cryptographers have proposed a hash algorithm that they dubbed MD6.
NIST plans to initially do a completeness check to make certain that every entry satisfies the requirements set by the government. The resulting list of submissions will be posted online by the end of the year, Burr said. In early 2009, the agency will hold a cryptography workshop to discuss the submissions. A second workshop will likely be scheduled in early 2010, by which time NIST aims to have the field whittled down to 15 algorithms, Burr said.
"If some of the proposals are pretty sketchy, then they are not going to make it compared with the proposals that explain why they did this or why they did that," Burr said. "It may be that we will decide that we want to do an earlier cut, so people can concentrate on the likely winners."
While the submissions have to gain check marks for NIST's requirements, the most significant gauntlet will be the scrutiny of other researchers.
"Over the next couple of years, a favorite occupation for grad students will likely be to break these algorithms," Burr said. "People are going to be writing papers about attacks on various hash functions. It's all fair game."
Burr believes that at least 40 teams will likely submit proposals, more than double the number of submissions to the contest to determine the Advanced Encryption Standard (AES). BT Counterpane's Schneier, a noted cryptographer, estimates that the field of submissions might rise to nearly 80, if interested amateur cryptographers make proposals as well.
More proposals is not necessarily better, Schneier said.
"The large number of entries is worrisome, because what it will do is dilute our (the community of cryptographers') analytical efforts," Schneier said. "The more there are, the less scrutiny that the proposals will get."
Yet, the cryptographers will have a lot of time to find the cracks in the hashing functions. The competition is a marathon, not a sprint, stressed Burr.
"SHA-3 is not going to be available, realistically, for another decade," he said. "If you look how long it is going to take us to pick it and how long it will take to make its way into products, well, we have been telling people not to wait on SHA-3."
If you have tips or insights on this topic, please contact SecurityFocus.
This article originally appeared in Security Focus.
Copyright © 2008, SecurityFocus