August 09, 2007

The Uneasy Ride on the Cryptography Bandwaggon

It is good to look beyond the basics and address the systemic aspects of failure. Neal Koblitz, who had something to do with the invention of ECC, describes in a forthcoming paper two bandwaggons that cryptographers have leaped on:

Koblitz describes two pernicious effects of this mixing of the two fields. One he calls the "bandwagon effect", in which mathematicians have distorted their research grant proposals in an effort to appeal to funding entities like the National Security Agency.

The other is the effort by various cryptographers to add an aura of reliability to their cryptographic systems by claiming the systems are "provably" secure---that is, by claiming there exists an ironclad mathematical proof of the system's security. Koblitz and a colleague have written several papers critiquing claims of "provable security", and he describes the heated and sometimes bizarre reactions that greeted their critique.

We've seen both those. Certainly the first is widespread.

What makes the second so interesting is that it wouldn't work in any other field, it is so hard for someone to knock down, and the concept has done so much damage that we now also write papers about why this error is so prevalent. C.f., Pareto-secure was an attempt on my part to explain just where "probably-secure" takes us, positively, and where the limitations are.

And, no, there is no connection to the photo. I just want one and I'm not at CCC to steal one...

Posted by iang at August 9, 2007 09:14 PM | TrackBack

In the meantime the paper is published:

Posted by: Twan van der Schoot at August 10, 2007 05:12 AM

Two guys credited with ecc in 1985 were Koblitz at univ. wash and Victor Miller at ykt (math dept). Coppersmith (lucifer, des) was also in the math dept at ykt.

During period that Koblitz and Miller were doing the ecc related work, i had offices in bldg. 28 (sjr) and bldg. 29 (vlsi lab) on the west coast ... but reported to ykt on the east coast ... which met x-cntry flts a couple times or more a month and interacted. i had interactions the people in the math dept (coppersmith, koblitz, etc) on various crypto stuff i was doing.

misc. old related to various kinds of crypto

misc references:,home,ecc_tut_1_0,290660,sid14_gci784941,00.html

Posted by: Lynn at August 10, 2007 11:25 AM
Post a comment

Remember personal info?

Hit preview to see your comment as it would be displayed.