all i know is if i browse to http://the-ip-of-my-site
i hit the vhost that matches the name the-ip-of-my-site
not the default vhost for 'empty hostname'

so i know that no http1.1 browser seen yet sends empty host header when 
browsing to an ip

(as the empty/unmatched host header vhost responds with a page offering http1.0 
links to all the sites, and http1.0 compatible links to download http1.1 
compatible clients (amazingly few browsers can be downloaded with a http1.0 
client btw) (tested with ncsa mosaic v1 ish)

and gets extremely few hits ever

but the ip as name vhost sees hundreds of hits an hour from 
probes/bots/attackers and logs/redirects or blacklists them based on attack 
type seen



At 07:34 29/03/2017  Wednesday, Niklas Keller wrote:

>From RFC 2616:
>
>A client MUST include a Host header field in all HTTP/1.1 request messages . 
>If the requested URI does not include an Internet host name for the service 
>being requested, then the Host header field MUST be given with an empty value.
>
>I guess an IP address is not considered an Internet host name and thus the 
>Host header must be empty. 
>
>Regards, Niklas 
>
>Richard Barnes <[email protected]> schrieb am Di., 28. März 2017, 21:13:
>On Tue, Mar 28, 2017 at 3:08 PM, Ilari Liusvaara 
><<mailto:[email protected]>[email protected]> wrote:
>On Mon, Mar 27, 2017 at 07:28:53PM -0400, Richard Barnes wrote:
>> Thanks, Roland.  Interesting draft.
>>
>> Couple of first reactions:
>>
>> - Why use the target of the PTR instead of just provisioning the TXT record
>> directly in the reverse DNS.  (Is there some restriction in the spec for
>> reverse DNS that says it's only PTR?)Â  It seems like by using the PTR
>> target, your security analysis gets much more complicated.
>
>Reading proposed IP validation rules in CABForum, using PTR target would
>certainly qualify, whereas putting TXT record on the name probably does
>not.
>
>The proposed rules have nothing on DNS apart from PTR lookups.
>
>(The current rules have "other equivalent method" clause.)
>
>So my reading is, that one has to do the PTR lookup in order to remain
>compliant with BRs after the proposed tighening of IP validation goes
>through (I expect it to).
>
>> - For the re-use of "http-01", you should probably specify the contents of
>> the Host header.  (Main ACME should probably clarify that for DNS, if it's
>> not clear already.)
>
>Reading the proposed and current rules, the challenge is either invoked
>on IP address itself, or on reverse-DNS lookup of it. So if you don't
>do rDNS lookup, then the validation is invoked on IP address itself,
>which I think leads to empty Host-header.
>
>
>Au contraire!
>
>rbarnes$ nc -l 8080 >/dev/null &
>rbarnes$ curl -v <http://127.0.0.1:8080>http://127.0.0.1:8080
>* Rebuilt URL to: <http://127.0.0.1:8080/>http://127.0.0.1:8080/
>* Â  Trying 127.0.0.1...
>* TCP_NODELAY set
>* Connected to 127.0.0.1 (127.0.0.1) port 8080 (#0)
>> GET / HTTP/1.1
>> Host: <http://127.0.0.1:8080>127.0.0.1:8080
>> User-Agent: curl/7.51.0
>> Accept: */*
>>Â 
>
>Guess this is good evidence that we should document this :)
>Â 
>
>TLS-SNI-02 would work as usual (the proposed rules explicitly list
>method 10, which is what TLS-SNI-02 falls under). Existing rules don't
>say anything (but again, equivalent method clause exists).
>
>
>
>
>
>-Ilari
>
>_______________________________________________
>Acme mailing list
><mailto:[email protected]>[email protected]
>https://www.ietf.org/mailman/listinfo/acme
>
>_______________________________________________
>Acme mailing list
>[email protected]
>https://www.ietf.org/mailman/listinfo/acme

_______________________________________________
Acme mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/acme

Reply via email to