Isn’t this a chicken vs egg problem?
Apparently:
> "The DOH server is given with a host name that itself needs to be resolved. This initial resolve needs to be done by the native resolver before DOH kicks in."
DNS over HTTPS
11–20 of 195 posts
Re: DNS over HTTPS
#12To understand DNS, you must first understand DNS.
DNS is one of those things I have to reference every time. Definitely feels like this.
Re: DNS over HTTPS
#13> for privacy, performance and security. I understand the privacy and security aspects. But I am wondering - how can DNS over HTTPS be more performant in the case of curl commands? A browser could probably persist the connection to the resolver and issue several requests together, but with a single curl command surely there's the overhead of initiating the first DNS resolve, the HTTPS connection, the second DNS resol…
cURL supports it. See https://ec.haxx.se/usingcurl-persist.html
If your question is how to reuse the connection pool for, say, 1000 DNS queries from one machine, curl already uses a connection pool. Next, to use fewer TCP connections and get faster response, my first proposal is HTTP/2 which supports TCP multiplexing such that multiple requests can be issued over a single TCP connection. Every response is return asynchronously, we can achieve nonblocking.
cURL also supports HTTP/2.
Speaking of word: it seems author of curl project prefers “curl” over “cURL”? Worth asking.
Re: DNS over HTTPS
#14> for privacy, performance and security. I understand the privacy and security aspects. But I am wondering - how can DNS over HTTPS be more performant in the case of curl commands? A browser could probably persist the connection to the resolver and issue several requests together, but with a single curl command surely there's the overhead of initiating the first DNS resolve, the HTTPS connection, the second DNS resol…
Ideally the browser and curl only use a local resolver, on your machine or router, which in turn can use a steadily alive or even multiple alive connections to the resolver.
Re: DNS over HTTPS
#15> for privacy, performance and security. I understand the privacy and security aspects. But I am wondering - how can DNS over HTTPS be more performant in the case of curl commands? A browser could probably persist the connection to the resolver and issue several requests together, but with a single curl command surely there's the overhead of initiating the first DNS resolve, the HTTPS connection, the second DNS resol…
cURL supports it. See https://ec.haxx.se/usingcurl-persist.html If your question is how to reuse the connection pool for, say, 1000 DNS queries from one machine, curl already uses a connection pool. Next, to use fewer TCP connections and get faster response, my first proposal is HTTP/2 which supports TCP multiplexing such that multiple requests can be issued over a single TCP connection. Every response is return asyn…
From what I understand the "cURL" format of the word is popular mostly with users of the PHP bindings. "curl" is what the project uses, but I'm not sure they really care what other people refer to it as :)
Re: DNS over HTTPS
#16Where does the privacy advantage come from? Once you resolve a hostname privately, don't you still need to use its IP address publicly for your traffic to be routed there?
Re: DNS over HTTPS
#17Isn’t this a chicken vs egg problem?
Just stick the address of the DNS resolver in /etc/hosts.
Re: DNS over HTTPS
#18> for privacy, performance and security. I understand the privacy and security aspects. But I am wondering - how can DNS over HTTPS be more performant in the case of curl commands? A browser could probably persist the connection to the resolver and issue several requests together, but with a single curl command surely there's the overhead of initiating the first DNS resolve, the HTTPS connection, the second DNS resol…
cURL supports it. See https://ec.haxx.se/usingcurl-persist.html If your question is how to reuse the connection pool for, say, 1000 DNS queries from one machine, curl already uses a connection pool. Next, to use fewer TCP connections and get faster response, my first proposal is HTTP/2 which supports TCP multiplexing such that multiple requests can be issued over a single TCP connection. Every response is return asyn…
> Every response is return asynchronously, we can achieve nonblocking.
same goes for UDP which DNS runs over normally.
Like the original commenter, I fail to see how stacking DNS over HTTP over TLS over TCP over IP is going to be faster than running DNS over UDP over IP.
Also, you still need to resolve the address of your DNS-over-HTTP server, which you'd probably still do over traditional DNS-over-UDP.
This feels like doing nothing but adding complexity for zero gain.
Re: DNS over HTTPS
#19Earlier quoted context omitted.
cURL supports it. See https://ec.haxx.se/usingcurl-persist.html If your question is how to reuse the connection pool for, say, 1000 DNS queries from one machine, curl already uses a connection pool. Next, to use fewer TCP connections and get faster response, my first proposal is HTTP/2 which supports TCP multiplexing such that multiple requests can be issued over a single TCP connection. Every response is return asyn…
From what I understand the "cURL" format of the word is popular mostly with users of the PHP bindings. "curl" is what the project uses, but I'm not sure they really care what other people refer to it as :)
I think it depends on whether you are referring to the project or the executable [1]: "cURL is the name of the project."
Re: DNS over HTTPS
#20So what use does this have?
DNSSEC already gives us validation of the records, and thanks to SNI, this doesn't give us any privacy.
It is more complex, more centralized, and ends up slower than using actual DNS, and doesn't seem to provide any benefits.
Am I missing something?