Ahhh, leveraging the fact that the secure hashes are written in C
would definitely improve things.

Do you have a link to any info about the S3 hash problem?  A quick
search didn't turn anything up.

On Oct 2, 3:28 pm, yejun <[EMAIL PROTECTED]> wrote:
> The problem is that a hand written python hash function not likely
> outperform secure hash function implemented in c.
> This kind attack actually happened once on amazon's s3 service, which
> caused about 2 hours partial service interruption.
>
> On Oct 2, 4:43 pm, Tony Arkles <[EMAIL PROTECTED]> wrote:
>
> > I don't think a *secure* hash function is necessary here (nor do I
> > think it desirable, due to speed).  By the pigeonhole principle, *all*
> > hashes have collisions; it's just a matter of a) how likely it is to
> > happen, and b) how much of an impact will it have.
>
> > For doing a hashtable implementation (which is basically what we're
> > talking about here), it seems to me that it's much more important to
> > have a fast hash than it is to have a secure hash.  In theory an
> > attacker could submit a huge pile of specially-constructed URLs that
> > all had the same hash (such that the lookups would be linear instead
> > of constant), but I'm much more concerned about minimizing my CPU
> > burder on all requests :)
>
> > On Oct 2, 1:09 pm, yejun <[EMAIL PROTECTED]> wrote:
>
> > > On Oct 2, 2:31 pm, Bill <[EMAIL PROTECTED]> wrote:
>
> > > > Seems like the SHA hashes are overkill.
>
> > > Collisions have been found for both sha0 and md5.
> > > I guess the minimal usable hash function is sha1 for now. In python
> > > hashlib uses openssl which should be reasonable fast.
>
>
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"Google App Engine" group.
To post to this group, send email to google-appengine@googlegroups.com
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at 
http://groups.google.com/group/google-appengine?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to