Shibboleth and SHA-256 ?

[Ian Young]

On 8 May 2013, at 21:42, "Cantor, Scott" <cantor.2 at osu.edu> wrote:

> Ian has done a lot more investigating and planning for this than I have, and I believe the UK Federation has a tentative date at least for moving to SHA-256 signed metadata.

Yes, speaking with my UKf architect hat on, we are planning to move to an RSA+SHA-256 signature profile for production metadata before the end of the year.  Details are in our Federation Technical Specifications:

http://www.ukfederation.org.uk/library/uploads/Documents/federation-technical-specifications.pdf

or

http://bit.ly/Q7M9T2l

See section 4.5.5.

Today, our test and export aggregates are signed in this way, and we haven't seen any problems yet.  I'm hoping to build up some body of compatibility information before we switch the production aggregate over, and we have provisions for a transitional period  during which we will provide a fallback for people who turn out to have issues.  How long we need that for will depend on how many problems come out of the woodwork.

If anyone has specific information about software+platform combinations they know *cannot* validate SHA-256 signatures, I'd appreciate a heads-up.  If you have a combination that you *suspect* may be problematic, you can still let me know and we can work through the test.

I think getting a solid handle on this will help us (speaking now with my Shibboleth team hat on) make the right decision about the defaults for *signing* in the V3 IdP.  The NIST recommendations are pretty clear, though, that they regard RSA+SHA-1 signatures as "unsafe at any speed" from the end of 2013, so there's going to be significant pressure to go in that direction.  Accurate algorithm agility metadata may be one way to mitigate the downsides of that, but we're into another discussion now.

	-- Ian



-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 4813 bytes
Desc: not available
Url : http://shibboleth.net/pipermail/users/attachments/20130508/c6d7ea56/attachment-0001.bin