RE: [Cfrg] Request For Opinions

[David Mcgrew <[e-mail]>]

Jon,

thanks for clarifying.  I have a couple of questions.

The work that I'd cited was aimed at asymmetric authentication of packet 
flows.  One of the goals of this work is to limit the amount of data that a 
receiver needs to buffer, for obvious networking reasons.  However, I 
expect that in your case, buffering is not an issue.  Since we expect to be 
able to provide long-term nonrepudiation on the data, we're storing it 
somewhere :-)   So if there were no packet loss, then we could hash all of 
the packets in a flow, then sign it once its over.  You mention that a 
sequence number is associated with the packets in the flow, so I expect 
that there's a well-defined session.  (I'm curious here.  What is the 
protocol, and how is the 'session' defined?)   The sequence number would 
make the order of the packets fed into the hash unambiguous, so it seems to 
me that the packet loss is the only complication.  Thinking out loud, this 
would seem to be a problem if the point in the network where the 
nonrepudiation service were applied was different from the point that did 
the storage.  I expect that this is the case, since otherwise we could just 
use a trusted storage device and go home early that day :-)

David

At 10:09 AM 5/8/2003 -0400, [e-mail] wrote:
>Non-repudiation is a requirement, and going back to the source to validate
>the authentication has been ruled out.
>
>Putting these together with my weak understanding of the Gennaro-Rohatgi
>stream method, it seemed that scheme didn't work (had to be able to go back
>to the source, or have the stream stored ahead of time).
>
>Also the ruling out of going back to the source of the data for validation
>is what keeps us from using a nice fast keyed hash approach :(
>
>I don't think, however, taking the authentication out-of-band has been
>considered.  Not sure what it buys, but worth looking into.
>
>Thanks again,
>Jon
>
>-----Original Message-----
>From: Hallam-Baker, Phillip [mailto:[e-mail]]
>Sent: Thursday, May 08, 2003 9:52 AM
>To: 'canetti'; Hallam-Baker, Phillip
>Cc: 'David Mcgrew'; [e-mail]; [e-mail]
>Subject: RE: [Cfrg] Request For Opinions
>
>
>Ran,
>
>         I think you are probably being unkind to them, they have a solution
>for the infinite case, the scheme I suggest is actually a special case of a
>scheme they describe in the introduction. They also point to some pretty
>good prior art - the java libraries use this technique.
>
>http://citeseer.nj.nec.com/gennaro97how.html
>
>         The article also proposes a solution to the infinite streams case.
>
>
>         One point to consider is that the bar to an acceptable solution is
>pretty high since it has to address a problem for which HMAC is not
>acceptable (a hard non-repudiation requirement) and for which custom
>hardware is not acceptable.
>
>         One issue that I did think of pushing on is the question of whether
>true non-repudiation is the objective or whether it suffices to have an
>authentication metric that does not depend on a shared key across a
>population.
>
>
>         If I had an order to do streaming video over multicast the way I
>would do it is with two independent streams for each user. The first stream
>would be the shared multicast stream. The second stream would be a
>per-subscriber stream that contained the authentication data.
>
>         I would encapsulate the video feed in something like DIME and put a
>message digest (SHA1) at the end of each block.
>
>         Then I would send to each subscriber a UDP packet containing HMAC
>(MD, k) where k is the session key established between the user and the
>service.
>
>         The advantage of DIME is that you can separate the authentication
>channel or do it in band as you choose.
>
>         We really need a replacement for the current video-conference
>protocols. the H.3??? scheme stops at every firewall I know. So there is no
>way that most corporate users can videoconference from home. So there might
>be an opportunity to drop this type of idea into the same scheme.
>
>
>         Some other thoughts
>
>- could use some sort of scheme that was symmetric key based but each party
>only had a subset of the total keying material. Essentially a CDMA type
>idea, each party has a unique subset of the keyspace.
>
>So there are 8 keys
>
>A has the key map  10101010
>B has the key map  11001100
>C has the key map  11110000
>
>So A can validate the key stream but cannot generate the key stream unless
>she colludes with enough other parties to guarantee she has all the keys.
>
>
>To make the process harder there could be a much larger pool of keys than is
>used in each sample. The key maps would change from one step to another so
>that the key maps for block n would be different to those for block m.
>
>Anyway back to fighting spam, you know the simple stuff.
>
>                 Phill
>
>
> > -----Original Message-----
> > From: canetti [mailto:[e-mail]]
> > Sent: Thursday, May 08, 2003 9:04 AM
> > To: Hallam-Baker, Phillip
> > Cc: 'David Mcgrew'; [e-mail]; [e-mail]
> > Subject: RE: [Cfrg] Request For Opinions
> >
> >
> >
> > Phill,
> > This seems to me reminisent of the Gennaro-Rohatgi stream
> > signatures from
> > Crypto 97.
> >
> > Ran
> >
> >
> > On Thu, 8 May 2003, Hallam-Baker, Phillip wrote:
> >
> > > Is the data stored data or being generated on the fly?
> > >
> > > If the data is stored you could sign the data as follows:
> > >
> > > Divide the stream into n blocks, say 100 blocks of 8K
> > >
> > > Compute H(n)  the hash of block n,
> > > Compute H(n-1) the hash of block n-1 + H(n)
> > > and so on
> > > Compute H(1), the hash of block 1 + H(2)
> > >
> > > Then sign H(1) and put it at the START of the stream.
> > > Next send H(1), block 1
> > > H(2), block 2 etc...
> > >
> > >
> > > This lets you check the integrity of the stream as you read
> > the chunks, with
> > > the ability to verify each chunk as you read.
> > >
> > > You cannot of course sign an infinite stream unless you
> > have a buffer of
> > > infinite length.
> > >
> > >
> > > I have no idea if anyone has thought of this before
> > (probably have it seems
> > > obvious enough) or if it is patented or anything.
> > >
> > >         Phill
> > >
> > >
> > >
> > > -----Original Message-----
> > > From: David Mcgrew [mailto:[e-mail]]
> > > Sent: Monday, May 05, 2003 2:30 PM
> > > To: [e-mail]
> > > Cc: [e-mail]
> > > Subject: Re: [Cfrg] Request For Opinions
> > >
> > >
> > > Jon,
> > >
> > > it seems like you have a viable scheme.  You might be
> > interested in the
> > > efficient methods for signing of packet streams that have
> > been discussed in
> > > the IRTF, starting with the Secure Multicast/Group (SMUG) RG (
> > > http://www.securemulticast.org/smug-index.htm
> > > <http://www.securemulticast.org/smug-index.htm> ), which is
> > now defunct but
> > > has spawned both the IRTF Group Security (GSEC) RG and the
> > IETF MSEC WG.
> > > There are a bunch of methods that amortize the work of
> > digital signatures
> > > across multiple packets.  I think the following list covers the main
> > > methods:
> > >
> > > How to Sign Digital Streams, Gennaro and Rohatgi, Crypto
> > '97, online at
> > > http://citeseer.nj.nec.com/gennaro97how.html
> > > <http://citeseer.nj.nec.com/gennaro97how.html> .
> > >
> > > Digital Signatures for Flows and Multicasts, Chung Kei Wong
> > and Simon S.
> > > Lam,Technical Report TR-98-15, May 31, 1998; revised, June
> > 14, 1999, in
> > > IEEE/ACM Transactions on Networking, August 1999.
> > > http://www.cs.utexas.edu/users/lam/Vita/IEEE/WongLam99.pdf
> > > <http://www.cs.utexas.edu/users/lam/Vita/IEEE/WongLam99.pdf>
> > >
> > > Authenticating Streamed Data in the Presence of Random
> > Packet Loss, Golle
> > > and Modadugu, NDSS '00, online at
> > > http://crypto.stanford.edu/~nagendra/projects/StreamAuth/auth.pdf
> > > <http://crypto.stanford.edu/~nagendra/projects/StreamAuth/auth.pdf>
> > > Similar but independent work was done by Perrig, Song and
> > Canetti, IIRC,
> > > though I can't find a link for it at present.
> > >
> > > The Use of RSA Signatures within ESP and AH, Brian Weis,
> > >
> > http://www.ietf.org/internet-drafts/draft-bew-ipsec-signatures-00.txt
> > >
> > <http://www.ietf.org/internet-drafts/draft-bew-ipsec-signature> s-00.txt> .
> > > This method could be used with one of the standard methods
> > for accerating
> > > public key signing or verifying, though it's not explicitly
> > discussed in the
> > > document.
> > >
> > > A related but probably not applicable work is "TESLA:
> > Multicast Source
> > > Authentication Transform Specification", Perrig, Canetti, Whillock,
> > >
> > http://www.ietf.org/internet-drafts/draft-ietf-msec-tesla-spec-00.txt
> > >
> > <http://www.ietf.org/internet-drafts/draft-ietf-msec-tesla-spe> c-00.txt> .
> > >
> > > Which of these solutions is the best for any given
> > application depends on a
> > > bunch of factors, such as the possibility loss and reorder
> > of packets,
> > > whether or not it is acceptable to buffer packets on the
> > sender or on the
> > > receiver, and the amount of data that can safely be added
> > to each packet.
> > >
> > > Hope this help,
> > >
> > > David
> > >
> > >
> > > At 11:52 AM 5/5/2003 -0400, [e-mail] wrote:
> > >
> > >
> > > Hey all,
> > >
> > > I'm in an ETSI working group and we are trying to find a
> > way to provide
> > > integrity checks along with non-repudiation on a
> > theoretically infinite
> > > amount of data delivered over a network at 10's to 100's Mbps.
> > > Crypto-hardware is not a reasonable expectation and the solution is
> > > extremely cost sensitive.
> > >
> > > The data link will be sending PDU's an an as need basis,
> > with each PDU being
> > > tagged with a sequence number.  At some point later in time
> > > (days/weeks/years) we need to be able to prove the
> > integrity of the PDU's as
> > > well as the authenticity.
> > >
> > > The following is what we have come up with and I'm
> > wondering if anyone has
> > > anything to add, remove, or change that would make a better
> > solution.  The
> > > idea is to get some of the authentication strength of
> > DSS/DSA without the
> > > full strength or the full cost (in processing time).
> > >
> > > Thanks in advance for any input,
> > > Jon
> > >
> > > Periodically, hashes over the data PDUs will be inserted
> > into the data
> > > stream. A hash will only be created if at least one PDU
> > packet was sent
> > > since the last hash was sent or since startup.
> > >
> > > Every <n1> PDU packets or when <t1> seconds have passed a
> > MD5 hash is
> > > generated over all PDU
> > > packets sent since the last MD5 hash was sent or since
> > startup. The MD5 hash
> > > is sent in a HashPDU, where the HashSequenceNr increments
> > for every hash
> > > sent for this intercept. The array PDUSequenceNrsIncluded
> > contains the
> > > sequence number of every data PDU that was included in the hash.
> > >
> > > NOTE: The values for n1 and t1 are configurable.
> > >
> > > Periodically, a hash over the hashes specified above and a
> > signature of that
> > > hash, will be inserted into the  the data stream. The
> > signed hashes allows
> > > authentication of the sender and to verfify the integrity
> > of the recevied
> > > MD5 hashes.
> > >
> > > Every <n2> hashes or when <t2> seconds have passed, a MD5
> > hash is generated
> > > over all hashes sent since the last signed hash was sent or
> > since startup
> > > and that MD5 hash is signed
> > > using DSS/DSA. The MD5 hash and corresponding DSS/DSA
> > signature are sent
> > > over in a SignedHashPDU, where the SignedHashSequenceNr
> > increments for every
> > > signed hash sent. The array HashSequenceNrsIncluded
> > contains the sequence
> > > number of every HashPDU that
> > > was included in the signed hash.
> > >
> > > NOTE: The values for n2 and t2 are configurable.
> > >
> > > NOTE: The distribution of the DSS/DSA public key is outside
> > the scope of
> > > this specification.
> > >
> > > _______________________________________________
> > > Cfrg mailing list
> > > [e-mail]
> > > https://www1.ietf.org/mailman/listinfo/cfrg
> > > <https://www1.ietf.org/mailman/listinfo/cfrg>
> > >
> > > _______________________________________________ Cfrg mailing list
> > > [e-mail] https://www1.ietf.org/mailman/listinfo/cfrg
> > >
> > >
> >

_______________________________________________
Cfrg mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/cfrg