Re: Some possible attacks on SHA1

On 10/05/2012 04:35 PM, Melvin Carvalho wrote:
> This article shows that attacks could be feasible by 2018 
> http://www.schneier.com/blog/archives/2012/10/when_will_we_se.html

Thanks for the heads-up Melvin. We examined the attack, and while we
don't agree with Schneier's assertion about the financial cost of using
an AWS-like system to mount an attack on SHA-1, we do agree that the
possibility exists within the next decade.

This affects the PaySwarm specs, specifically the digital signature
algorithm for signed JSON-LD messages. We explored the idea that we
could greatly reduce the SHA-1 attack by injecting the length of the
message in the generation of the digital signature, but instead chose to
upgrade the spec requirements to SHA-256.

SHA-256 has no known theoretical attack at present, nor is a brute-force
attack on the algorithm known to exist that can be accomplished in the
near term (less than 10 years into the future). As with all things
crypto-related, this may change tomorrow, but SHA-256 seems to be the
right solution today. SHA-3 is too new, but I expect that we will
eventually end up using it.

Dave Longley has already committed the changes to the production
PaySwarm code. We'll push the changes to the dev.payswarm.com site soon.
The PHP WordPress PaySwarm client was updated today:

https://github.com/digitalbazaar/payswarm-wordpress/commit/0e0be20f20508998d04c95dc4a3009cd2e176a01

as well as the JavaScript PaySwarm client:

https://github.com/digitalbazaar/payswarm.js/commit/b8f2ce880c8858b27fad3d572350a22243d85aa3

-- manu

-- 
Manu Sporny (skype: msporny, twitter: manusporny)
President/CEO - Digital Bazaar, Inc.
blog: The Problem with RDF and Nuclear Power
http://manu.sporny.org/2012/nuclear-rdf/

Received on Sunday, 7 October 2012 03:11:43 UTC