[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: uniqueness of Message IDs and related issues

>   The MUST clause is not hyperbole. It states that the
> message ID of the
> notify MUST NOT be the same as the message ID of the Quick
> Mode exchange
> that generated the error. This is because, as you noted, the
> message ID
> is used to demultiplex the exchanges that two peers may be engaging in
> at any one moment-- this can be more than 1. This verbage resolved an
> interoperability issue some time ago.

Sure, but I was actually refering to this quote:

   The message ID used for all non-Phase 1 exchanges MUST be pseudo-
   randomly generated using a strong random number generator.

I think that's a bit of hyperbole. As you mentioned, the important thing is
that the Message ID is not used for two exchanges which occur in temporal
proximity (if you'll excuse the Star Trek-ish terminology).

>   Also, if the message ID has no cryptographic significance
> then it does
> not need to be based off a "truely random seed".

Okay, sure. The point of my original message was that you don't need to keep
adding entropy for every Message ID you generate. You just need a stream of
unpredictable data; it doesn't have to be truly random. (and to defeat an
adversary who can guess your entire state you need a truly random seed)

Andrew
--------------------------------------
Beauty with out truth is insubstantial.
Truth without beauty is unbearable.


> -----Original Message-----
> From: Dan Harkins [mailto:dharkins@cips.nokia.com]
> Sent: Thursday, June 22, 2000 1:26 PM
> To: andrew.krywaniuk@alcatel.com
> Cc: 'IPsec List'
> Subject: Re: uniqueness of Message IDs and related issues
>
>
>   OK, I'll disagree with some of it.
>
>   The MUST clause is not hyperbole. It states that the
> message ID of the
> notify MUST NOT be the same as the message ID of the Quick
> Mode exchange
> that generated the error. This is because, as you noted, the
> message ID
> is used to demultiplex the exchanges that two peers may be engaging in
> at any one moment-- this can be more than 1. This verbage resolved an
> interoperability issue some time ago.
>
>   Also, if the message ID has no cryptographic significance
> then it does
> not need to be based off a "truely random seed".
>
>   Dan.
>
> On Thu, 22 Jun 2000 12:44:50 EDT you wrote
> >
> > I believe the MUST clause in the draft is a slight bit of hyperbole.
> >
> > Anyone want to add to/disagree with any of this?
> >
> > Andrew
>
References: