What do we mean by a replay attack? about python-tuf HOT 5 CLOSED

If in step 6, the attacker presents release 2, this is a freeze attack, not
a replay attack. The client doesn't see step 5 so it can be ignored.
Replay attacks are all about what the client sees.

A replay is either (using your steps from above):

1, 2, 3, 4, and attacker tries to provide release 1 to the client (your
step 2)

or

1, 3, 4, and attacker tries to provide release 1 to the client (your step 2)

from python-tuf.

Yes, agreed. What we have in place now seems to me more like an arbitrary package attack. Perhaps we should create a wiki page where we specify the attacks in more precise terms like the example above.

from python-tuf.

What we have in place where? Do you mean Kon's code?

Kon: I think what you had when I last looked was right. Can you take a
look at Trishank's description here and let me know what case you're doing?

Thanks,
Justin

On Fri, Mar 15, 2013 at 9:13 AM, TKK [email protected] wrote:

Yes, agreed. What we have in place now seems to me more like an arbitrary
package attack. Perhaps we should create a wiki page where we specify the
attacks in more precise terms like the example above.

—
Reply to this email directly or view it on GitHubhttps://github.com/akonst/tuf/issues/38#issuecomment-14959629
.

from python-tuf.

Sure, let's quickly go over it in the meeting today.

Sent from my iPad

On Mar 15, 2013, at 9:29 AM, JustinCappos [email protected] wrote:

What we have in place where? Do you mean Kon's code?

Kon: I think what you had when I last looked was right. Can you take a
look at Trishank's description here and let me know what case you're doing?

Thanks,
Justin

On Fri, Mar 15, 2013 at 9:13 AM, TKK [email protected] wrote:

Yes, agreed. What we have in place now seems to me more like an arbitrary
package attack. Perhaps we should create a wiki page where we specify the
attacks in more precise terms like the example above.

—
Reply to this email directly or view it on GitHubhttps://github.com/akonst/tuf/issues/38#issuecomment-14959629
.

—
Reply to this email directly or view it on GitHub.

from python-tuf.

Konst, feel free to close this issue if you think it has been addressed.

from python-tuf.