4 ms·
It's presented from a biased point of view. It doesn't discuss the problems with DoH.
by johnklos 7y ago
It's presented from a biased point of view. It doesn't discuss the problems with DoH.
- zamadatix 7y agoMost of the "problems" with DoH tend to be implementation based (e.g. applications individually resolving DNS or corporations assuming if they intercept UDP 53 they have secured connectivity) not necessarily problems with DoH. The main drawback of DoH itself is it moves DNS to a session based service which causes increases in workload for DNS infrastructure but to be fair the article does mention that so I'm not sure "it doesn't discuss the problems with DoH" is totally fair.
- dngray 7y agoI use my system resolvers on all my hosts, hand the IP to my unbound DNS server out and recurse over DoHs, that way everything is doing DoHs even if the underlying software doesn't implement it. DNSCrypt can do both it's protocol and DoHs. https://wiki.archlinux.org/index.php/Dnscrypt-proxy#Local_DNS_cache_configuration https://wiki.archlinux.org/index.php/Dnscrypt-proxy#Local_DN...
- LIV2 7y agoIt will be interesting to see how chaotic it will be for orgs that have split horizon DNS or rely on search domains to resolve things. I wonder how they're going to have to work with things like the Australian/NZ censorship which I believe is based on DNS?