13 Úno
2017
13 Úno
'17
17:51
Hi,
We would like to enable ECS responses in our Knot server in order to
indicate to recursives that they should send it in requests if they can. The
goal is gauge how widespread usage of this feature this is and to collect
some statistics on the client networks being transmitted. We don't have a
way of providing tailored responses yet (not sure how this would be done,
but that's another topic.) Is there a way of enabling ECS so that it's
included in responses, perhaps with a scope prefix-length of 0 (as
described in section 12.1 of draft-ietf-dnsop-edns-client-subnet-08)?
Thanks,
Chuck
13 Úno
13 Úno
20:09
Hi Chuck,
Currently Knot DNS doesn't support ECS. It's a planned but still
unfinished feature. However, it's possible to create a quite simple
patch
which could solve your problem.
Regards,
Daniel
On 2017-02-13 17:51, Chuck Musser wrote:
Hi,
We would like to enable ECS responses in our Knot server in order to
indicate to recursives that they should send it in requests if they
can. The
goal is gauge how widespread usage of this feature this is and to
collect
some statistics on the client networks being transmitted. We don't have
a
way of providing tailored responses yet (not sure how this would be
done,
but that's another topic.) Is there a way of enabling ECS so that it's
included in responses, perhaps with a scope prefix-length of 0 (as
described in section 12.1 of draft-ietf-dnsop-edns-client-subnet-08)?
Thanks,
Chuck
_______________________________________________
knot-dns-users mailing list
knot-dns-users(a)lists.nic.cz
https://lists.nic.cz/cgi-bin/mailman/listinfo/knot-dns-users
15 Úno
15 Úno
13:17
Hi Chuck,
I have some questions about what you exactly need to support.
Do you want to include ECS to every response or only to responses which
contains ECS in query?
Do you want to response with same Client Subnet used in query or always
use Client Subnet with prefix length of 0?
Regards,
Vita
Dne 13.2.2017 v 17:51 Chuck Musser napsal(a):
Hi,
We would like to enable ECS responses in our Knot server in order to
indicate to recursives that they should send it in requests if they can. The
goal is gauge how widespread usage of this feature this is and to collect
some statistics on the client networks being transmitted. We don't have a
way of providing tailored responses yet (not sure how this would be done,
but that's another topic.) Is there a way of enabling ECS so that it's
included in responses, perhaps with a scope prefix-length of 0 (as
described in section 12.1 of draft-ietf-dnsop-edns-client-subnet-08)?
Thanks,
Chuck
_______________________________________________
knot-dns-users mailing list
knot-dns-users(a)lists.nic.cz
https://lists.nic.cz/cgi-bin/mailman/listinfo/knot-dns-users
17 Úno
17 Úno
7:52
On Feb 15, 2017, at 4:17 AM, Vítězslav Kříž
<vitezslav.kriz(a)nic.cz> wrote:
Hi Chuck,
I have some questions about what you exactly need to support.
Do you want to include ECS to every response or only to responses which
contains ECS in query?
Only responses that contain it in the query. It’s probably
wise to only
send it in response to servers that appear able to use the information
and are likely to include ECS on subsequent requests.
Do you want to response with same Client Subnet used in query or always use Client Subnet
with prefix length of 0?
I think client subnet, with scope prefix length of zero.
The intent is to indicate that the
feature is supported but isn’t being used to (yet anyway) to tailor the response.
10:31
Hi,
On 17-02-17 07:52, Chuck Musser wrote:
I only see the need for ECS in knot if you can stick in a plugin that
can do something useful with the client address and update the scope
prefix length appropriately.
A generic plugin won't be particular useful because how to use the
client address to tailor the response is left unspecified and everybody
is likely to do it different anyway.
Best regards,
Matthijs
Chuck
_______________________________________________
knot-dns-users mailing list
knot-dns-users(a)lists.nic.cz
https://lists.nic.cz/cgi-bin/mailman/listinfo/knot-dns-users
On Feb 15, 2017, at 4:17 AM, Vítězslav Kříž
<vitezslav.kriz(a)nic.cz> wrote:
Hi Chuck,
I have some questions about what you exactly need to support.
Do you want to include ECS to every response or only to responses which
contains ECS in query?
Only responses that contain it in the query. It’s probably
wise to only
send it in response to servers that appear able to use the information
and are likely to include ECS on subsequent requests.
Do you want to response with same Client Subnet used in query or always use Client Subnet
with prefix length of 0?
I think client subnet, with scope prefix length of zero.
The intent is to indicate that the
feature is supported but isn’t being used to (yet anyway) to tailor the response.
12:49
Hi Chuck,
you can use branch ecs-pach:
https://gitlab.labs.nic.cz/labs/knot/tree/ecs-patch
https://gitlab.labs.nic.cz/labs/knot/commit/9fe76fe7dad761b05863e94cf36cf7c…
It responses with scope prefix length of zero only to query with ECS.
Regards,
Vita
Dne 17.2.2017 v 07:52 Chuck Musser napsal(a):
Chuck
On Feb 15, 2017, at 4:17 AM, Vítězslav Kříž
<vitezslav.kriz(a)nic.cz> wrote:
Hi Chuck,
I have some questions about what you exactly need to support.
Do you want to include ECS to every response or only to responses which
contains ECS in query?
Only responses that contain it in the query. It’s probably
wise to only
send it in response to servers that appear able to use the information
and are likely to include ECS on subsequent requests.
Do you want to response with same Client Subnet used in query or always use Client Subnet
with prefix length of 0?
I think client subnet, with scope prefix length of zero.
The intent is to indicate that the
feature is supported but isn’t being used to (yet anyway) to tailor the response.
2632
days inactive
2636
days old
5 comments
4 participants
participants (4)
-
Chuck Musser -
daniel.salzman＠nic.cz -
Matthijs Mekking -
Vítězslav Kříž