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

Re: header-swap isn't perfectly indistinguishable (was: problem in 3.2 "Replies")




 Len wrote:
>
> Um, I'm confused. The header size is fixed. The number of hops is
> variable. Unused header slots are padded. How does this reveal any
> information?

I'm sorry if I didn't make this clear in the message.  It is easy to confuse 
the issue.  There isn't anything in a normal non-swap message which reveals 
whether it is sender-anonymous or mutually-anonymous.  Nor is there anything 
in a swap message which reveals this.  But the fact of whether a message is a 
swap or a non-swap message reveals some probalistic information about whether 
the message belongs to a sender-anonymous or a mutually-anonymous chain.

This is because these two kinds of chains are of different lengths, but they 
each feature exactly one swap step, so the probability of a swap step occuring 
in an arbitrary hop is different for the different kinds of chain.

BTW, after I wrote the previous message it occurred to me that this might be 
fixed if mutually-anonymous messages featured two swap steps, just so that the 
ratio of swap steps to total hops would be the same as for sender-anonymous 
messages...

Regards,

Zooko

Zooko.Com -- Security and Distributed Systems Engineering