[Rpm-metadata] MD5 is vulnerable...
seth vidal
skvidal at phy.duke.edu
Sat Feb 19 16:24:35 UTC 2005
On Wed, 2005-02-16 at 23:29 -0600, Randy Zagar wrote:
>It's been known for at least a few months that there are
>block-transposition attacks that can cause collision problems with MD5.
>This means that it is theoretically possible to construct a new
>(possibly hostile) file that has the same MD5 checksum as a file that we
>trust...
>
>Bruce Schneier also reports on his weblog that SHA-1 has been
>compromised. The difficulty of cracking SHA-1 checksums has been
>reduced from "virtually impossible" to "extremely difficult".
>
>In light of this, I'd like to suggest that the xml metadata be modified
>to include BOTH the MD5 and SHA-1 checksums in the metadata files...
>Even if someone manages to compromise one signature, it'll be
>geometrically more difficult to compromise both signatures simultaneously...
not to make light of this but, umm, you're kidding, right?
-sv
More information about the Rpm-metadata
mailing list