RIPE NCC’s Post

View organization page for RIPE NCC

19,314 followers

🔒 DNS encryption is gaining momentum with proposed standards such as DoT, DoH, and DoQ protecting DNS exchanges from external observers. One key piece? Discovery of Designated Resolvers (DDR) — a mechanism that lets clients learn the encryption settings of recursive resolvers. For RIPE Labs, Yevheniya Nosyk looks at DDR in action and how it’s being deployed in the wild: https://lnkd.in/e2e_mpjg With contributions from Andrzej Duda and Maciej Korczynski

  • text
Tommy Jensen

Networking protocol technologist

1w

UPDATE: original comment remains here, but see my reply below. I really like the data analysis, thank you for publishing this. However, I have to be pedantic point out that this statement is misleading: "the great majority of resolvers designate one of the top 5 operators, " The act of designating only applies to an unencrypted resolver designating an encrypted resolver as its equivalent. This method (RFC 9462 section 4) does not permit a resolver to "designate one of the top" because the security check will fail: the destination resolver has to demonstrate control over the original IP address. The other mechanism we defined (section 5) you accurately describe early in the post, but I think is being referred to as designation in the quote above when it is just discovery. If you query 8.8.8[.]8 over DoH for one[.]one[.]one[.]one's encrypted DNS config using DDR, that isn't 8.8.8[.]8 designating to 1.1.1[.]1. This is broken down in the final paragraph of RFC 9462's introduction. In short: DDR doesn't enable centralization. Analyzing it may reveal DNS usage centralization (because most lookups of configuration are for only the big resolvers), but random resolvers cannot designate one of the big resolvers in their own place.

Like
Reply

To view or add a comment, sign in

Explore content categories