Re: [Long] A history of Netscape/MSIE problems

[Adam Back <aba@dcs.ex.ac.uk>]

Peter Guttmann <pgut001@cs.auckland.ac.nz> writes on cpunks:
> [...]  The reason for the 40-bit key and (according to RSADSI, the
> company that developed RC4) the reason why details on it were kept
> secret was that these conditions were required under an agreement
> between the Software Publishers Association (SPA) and the US
> government which gave special export status to the RC4 algorithm and
> a companion algorithm called RC2.

Hadn't heard that before, that the trade secret requirement was
imposed on RSADSI.  What was your source for that info, it is an
interesting assertion on the part of RSADSI, and I am intrigued.

> [reverse engineer of RC4...]
> The results were posted to mailing lists and the Internet [Anon
> 1994a].  Someone with a copy of BSAFE tested it against the real
> thing and verified that the two algorithms produced identical
> results [Rescorla 1994], and someone else checked with people who
> had seen the original RC4 code to make sure that it had been
> (legally) reverse-engineered rather than (illegally) copied [Anon
> 1994b].

Some people held that it had been a licensed holder of RC4 source who
had posted it in violation of the license agreement.  I think I recall
that Tim May, may be others, argued this nearer the time.

That the code looked different isn't of itself proof that it was or
wasn't reverse engineered; it is entirely plausible for the anonymous
poster (if it was a source license violation) to have gone to some
pains to obscure this fact, by changing the appearance and style of
the code.

> [RC4 key schedule biases...]

You ought to reference Andrew Roos paper [posted to the list, and
sci.crypt, at least] analysing key schedule biases in RC4.  Paul
Kocher posted a response (this was in sci.crypt) saying that he had
discovered the same biases while working for RSADSI, (at a time before
RC4 was revealed, or at least before RSADSI started discussing RC4
publically, a tacit admission by them that alleged RC4 was RC4)

> Further improvements to the attack were proposed.  

Andrew Roos brutessl code was special case optimised for SSL, he
precomputed part of the MD5 digest, and progressed through the key
space in an order chosen to maximise the amount of MD5 precomputation
that could be done.  Something of interest, perhaps.

> The attacks on RC4 are a prime example of a publicity attack.  They were
> carried out by volunteers using borrowed machine time, noone (apart from
> Netscapes stock prices) was harmed, 

Strangly (I'm not sure if anyone lost money due to this), I think
Netscapes prices hardly suffered, perhaps even improved slightly.
Could be due to the `any publicity is good publicity' syndrome.  There
was a *lot* of publicity, and Netscapes response in fixing the problem
was good.  Several US cypherpunks were tracking the stocks at the
time, and could probably verify this.

One omission: you didn't say anything about Paul Kocher's timing
attack on RSA, which I think affected Netscape servers, and was fixed
after his publicizing the attack.  Then you could discuss Ron Rivest's
blinding solution, and the time delay solution.

Otherwise, excellent.

Adam
--
#!/bin/perl -sp0777i<X+d*lMLa^*lN%0]dsXx++lMlN/dsM0<j]dsj
$/=unpack('H*',$_);$_=`echo 16dio\U$k"SK$/SM$n\EsN0p[lN*1
lK[d2%Sa2/d0$^Ixp"|dc`;s/\W//g;$_=pack('H*',/((..)*)$/)