5 ms·
The draft (the last version I've found is March 2009): http://tools.ietf.org/html/draft-rescorla-tls-extended-random-02 http://tools.ietf.org/html/draft-rescor
by acqq 13y ago
The draft (the last version I've found is March 2009):
http://tools.ietf.org/html/draft-rescorla-tls-extended-random-02 http://tools.ietf.org/html/draft-rescorla-tls-extended-rando...
(The first seems to be "Opaque PRF Inputs for TLS" from December 2006. http://tools.ietf.org/html/draft-rescorla-tls-opaque-prf-input-00 http://tools.ietf.org/html/draft-rescorla-tls-opaque-prf-inp... so the process lasted more than two years)
The relevant parts from the last:
TLS [I-D.ietf-tls-rfc4346-bis] and DTLS [RFC4347] use a 32-byte
"Random" value consisting of a 32-bit time value time and 28 randomly
generated bytes:
struct {
uint32 gmt_unix_time;
opaque random_bytes[28];
} Random;
The United States Department of Defense has requested a TLS mode
which allows the use of longer public randomness values for use with
high security level cipher suites like those specified in Suite B
[I-D.rescorla-tls-suiteb].
This document defines a new TLS extension called "extended_random".
The "extended_random" extension carried in a new TLS extension called
"ExtendedRandom".
struct {
opaque extended_random_value<0..2^16-1>;
} ExtendedRandom;
Note how they used the notation that casually looks smaller as in "there's 2 and 16 and 1" when they actually mean up to 65535 bytes (I've first thought it's fixed but it looks the length is a subject of negotiation).
Roughly 2300 times more than the default 28 bytes.
TPtacek says: "It's now thought to be risky to disclose too much state. But that was definitely not the attitude in place in the early 2000s --- more randomness was better."
I can't imagine that there was ever an attitude that it's not risky "to disclose too much state."
- tptacek 13y agoI'm not sure what you're trying to say here that my comment doesn't already say. It might be helpful to note that there are crypto people --- some of the authors on the recent Dual_EC exploit paper included --- that believe TLS already discloses too much random state, even without the TLS extension. I think it's uncontroversial in 2014 to say that any more state disclosure --- 2 bytes, 200 bytes, 16k bytes --- is probably not a good thing. It might be a little controversial to say this next thing, but I'll stand by it for the sake of an interesting argument: in 2005, when this discussion started, it was not known to be a bad thing to increase the amount of random state in the handshake.
- acqq 13y agoAs Thomas is not sure what I'm "trying to say here that" his "comment doesn't already say," he actually wrote: "All extended_random_value does is add, to the random_bytes value in the TLS handshake, another variable-length blob of random data." It sounds smallish "just another random-data blob." It's actually huge "up to additional 65365 bytes from the generator added to the default 28" before the encryption even starts. Thomas also wrote "The ostensible idea behind the extension is that there are protocols that want randomness proportional to key sizes." The biggest ECC key is 571 bits which fit in just 72 bytes and this thing provides up to 65535 bytes. Again, fully different impression. I quoted the draft to show the actual number. The numbers matter here. Hadn't I looked at the draft directly, I'd get the fully different impression only reading Thomas' comment.
- tptacek 13y agoI think you're missing the part where people can make credible arguments that twenty eight bytes of random state (the protocol as standardized) is already a problem. The difference between 10 and 60,000 extra bytes seems to make a big difference to you, and, fair enough I guess, but the underlying problem isn't the size of the random state. Read the Dual_EC exploit paper that just came out; the authors used minimal compute resources and standard TLS to exploit a model Dual_EC problem.
- acqq 13y agoAbout that (even a little more over 28 bytes makes the big change) we can agree: The paper http://dualec.org/DualECTLS.pdf http://dualec.org/DualECTLS.pdf shows that exploiting Dual EC takes hours (without additional bytes), but that RSA (intentionally or unintentionally) managed to adjust the BSAFE C implementation to enable especially easy exploits: in seconds vs more than an hour on the 16-CPU AMD cluster for other implementations (Table 1 in the PDF). The reason: "due to the use of the requested number of bytes rather than the number of remaining bytes after pulling from the cache, the concatenation of the 32-byte session ID and the 28 pseudorandom bytes in the server random always contains a full 30-byte output block and between one and 30 bytes of a subsequent block." So the 28 default bytes take hours, just adjusting that a little to push always continuous 58 bytes from the PRNG resulted in reducing the exploit time to seconds. The small number of bytes more, the big difference. Paper's conclusion: "Our analysis strongly suggests that, from an attacker's perspective, backdooring a PRNG should be combined not merely with influencing implementations to use the PRNG but also with influencing other details that secretly improve the exploitability of the PRNG." This is exactly why having the "extended_random_value" too was so convenient to the NSA.
- TrainedMonkey 13y agoAre there any details on how negotiation is handled? I think this may be innocuous by itself, however coupled with the software that requests large amount of randomness it is dangerous.
- tptacek 13y agoIt's not innocuous. There's no debate here about whether it's a good idea to have protocol participants expose extra random state on the wire. They shouldn't. As for details: we're talking about public Internet drafts.