Hi, Julien, Your example is correct. However, it is misleading. You were talking about you have found a collisional public-private key pair. In that case, no matter how strong the hash algorithms are, you can break the whole security system. Our assumption here is: based on a collision free public-private key pair, whether one hash algorithm with two results or two hash algorithms with two results are stronger. My choice is the latter.
Best regards, Sheng JIANG, Ph.D. IP Research Department, Networking Research Department, Network Product Line, Huawei Technologies Co. Ltd. *-----Original Message----- *From: julien laganier [mailto:[EMAIL PROTECTED] On *Behalf Of Julien Laganier *Sent: Wednesday, July 02, 2008 6:17 PM *To: Sheng Jiang *Cc: [email protected]; [EMAIL PROTECTED] *Subject: Re: [CGA-EXT] Hash Agility in SEND (Was: new version *of "Send Hash Threat Analysis") * *Hi Sheng, * *If I can generate a public-private key pair that yields an *arbitrary CGA (attack against your A below), then I can *impersonate that CGA's owner by simply signing with that *public-private key pair -- no need to break the digital *signature algorithm (your B). * *--julien * *On Wednesday 02 July 2008, Sheng Jiang wrote: *> Hi, Julien, *> *> I can agree with you that SEND and CGA may use the same hash *function. *> Actually, in the current protocol, they do use the same sha-1. *> However, we should also allow/support that they use different *> algorithms through we may not recommended it. *> *> I don't agree with your statement "the security of the resulting *> system can't be stronger than the weakest of its components". *> Actually, in many cases, it may be opposite: the strongest component *> decides the attacker's difficulty. Let's put it into a concrete *> example here, say, SEND uses algorithm B that stronger than *algorithm *> A in CGA and there is an attacker who can break algorithm A *but not B. *> Then, the attacker may produce a faked CGA option, but because he *> cannot produce a faked SEND signature option that use *algorithm B, he *> can still get through the SEND verification. When multiple *algorithms *> are use together, the attacker has to break all algorithms to break *> the security system. *> *> Best regards, *> *> Sheng *> *> *-----Original Message----- *> *From: [EMAIL PROTECTED] *> *[mailto:[EMAIL PROTECTED] On Behalf Of Julien Laganier *> *Sent: Tuesday, July 01, 2008 10:24 PM *> *To: [email protected]; [EMAIL PROTECTED] *> *Subject: [CGA-EXT] Hash Agility in SEND (Was: new version of *"Send *> Hash Threat Analysis") *> * *> * *> *Hello Ana, CSI'ers, *> * *> *I think I disagree with the paragraph of the draft copied *below my *> note. *> * *> *To me it seems OK to specify that SEND should use the same *hash *> function than CGA. Since they are used together to *provide *security, *> and since the security of the resulting *system can't be *stronger than *> the weakest of its components, *the maximum level of security can be *> reached by choosing *mechanisms (hash functions in this case) with *> similar security *strength for both CGA and SEND. *> * *> *Improving the security level of only one of the component **would not *> increase the overall security of the system. *> * *> *Has anybody an opinion on the topic? *> * *> *--julien *> * *> *> 4. Support for the hash agility in SeND *> *> *> While all of analyzed hash functions in SeND are theoretically *> *> affected by recent collision attacks, these attacks indicate *> the *> possibility of future real-world attacks. In order to *> *increase the *> *> future security of SeND, we suggest the support for the *hash and *> *> algorithm agility in SeND. *> *> *> *> The most effective and secure would be to bind the hash *> function *> option with something that can not be changed at all, *> like *> [rfc4982] does for CGA - encoding the hash function *> information into *> addresses. But, there is no possibilty *to do that *> in SeND. We could *> decide to use by default the same hash *function *> in SeND as *in CGA, but *> this solution is architecturally strange *> and it does not really *> increase the security since the difficulty *> for attackers remain to *> break one single hash function. *> Furthermore, it may even reduce the *> security level by providing *> more relevant information of the hash *> function. On the *other side, *> the use of two different hash *algorithms *> makes attacker's life *> harder. *> *> *> *> Another solution is to incorporate the hash function option *> into *> SeND message. By putting a new hash function option in SeND *> message *> before RSA Signature option, attacker will have to break *> both the *> signature and the hash input at the same time since the *> new option *> will be input field for the Digital Signature in RSA *> *Signature option. *> *> However, we can not avoid a downgrade attack totally because peer *> *> might be using just ND and not SeND. A completely safe *solution *> here *> does not exist. A new hash function option in SeND *message is *> a *> reasonable and the best solution for the hash algorithm agility *> *> support in SeND. *> *> *> *> Each implementation SHOULD use different hash or signature *> *> algorithms for each of the relevant fields (Key Hash *field, Digital *> *> Signature, PKIX signature algorithm). Since all *algorithms are in *> *> different procedures, making them the same does not make those *> *> procedures simpler, but making them different complicates *possible *> *> attacks. *> * *> * *> *On Tuesday 01 July 2008, Ana Kukec wrote: *> *> The new version of draft-kukec-csi-hash-threat is submitted. *> *Comments *> *> are welcome! *> *> *> *> Filename: draft-kukec-csi-hash-threat *> *> Revision: 02 *> *> Title: SeND Hash Threat Analysis *> *> Creation_date: 2008-07-01 *> *> WG ID: Independent Submission *> *> Number_of_pages: 15 *> *> *> *> Abstract: *> *> This document analysis the use of hashes in SeND, *possible *threats *> and *> the impact of recent attacks on hash functions used by SeND. *> *> Current SeND specification [rfc3971] uses SHA-1 [sha-1] hash *> *algorithm *> and PKIX certificates [rfc3280] and does not provide *> support for the *> hash algorithm agility. Based on previous *> analysis, this document *> suggests multiple hash support *that should *> be included in the SeND *> update specification. *> *> *> *> *> *> Ana *> *> *> *> _______________________________________________ *> *> CGA-EXT mailing list *> *> [email protected] *> *> https://www.ietf.org/mailman/listinfo/cga-ext *> * *> * *> * *> *_______________________________________________ *> *CGA-EXT mailing list *> [EMAIL PROTECTED] *> *https://www.ietf.org/mailman/listinfo/cga-ext *> * * * * _______________________________________________ CGA-EXT mailing list [email protected] https://www.ietf.org/mailman/listinfo/cga-ext
