1 |
On Thu, 28 Mar 2013 17:04:25 -0400 |
2 |
Michael Mol <mikemol@×××××.com> wrote: |
3 |
|
4 |
> > |
5 |
> >> listened to the dangers and even now simply redesigned DNSSEC. |
6 |
> > |
7 |
> > Or they could fudge it by making every request requiring padding |
8 |
> > larger than the response. Bandwidth would increase astronomically |
9 |
> > but amp attacks would have to find other avenues. |
10 |
> > |
11 |
> |
12 |
> Infeasible; the requester cannot know the size of the response in |
13 |
> advance. If a packet comes in, and the response is larger than the |
14 |
> request, is it really an amp packet, did the client not know, or is |
15 |
> the server misconfigured and not limiting the response data as much |
16 |
> as it could? |
17 |
|
18 |
I'm certainly not saying it's a good idea, hence the 'fudge' and 'making |
19 |
every request' which would mean non updateable clients or non updated |
20 |
routers (90%) needing special treatment. I'm sure there are probably |
21 |
other hurdles to it but it is certainly possible to make a request much |
22 |
larger than any potential response similar to the anti-spam system |
23 |
that makes creating a message take a lot of cpu and then only accepting |
24 |
messages from those that do (hsomething I think, only works too if all |
25 |
take part but would eliminate spam almost completely). |
26 |
|
27 |
However thinking about it, considering the want for dns to provide |
28 |
larger things like encryption keys, huge requests may be the best long |
29 |
term solution for a DNSSEC which seemingly refuses out of pride to add |
30 |
something like DNSCURVE to prevent spoofing. Similar to firewalls only |
31 |
sending a single syn ack (less than or equalise) |