Do not Associate Internet Centralization with DNS in the Absence of Facts
The last decade witnessed an ever increasing centralization of the Internet, which has triggered a grassroots movement toward Internet decentralization. However there seems an urgent need to understand the root cause of Internet centralization, so that all the decentralization efforts could lead to effective outcome.
One view that I have heard from many people at many places, IETF community included, is that DNS is a centralized system and therefore the decentralization efforts must find effective solutions to circumvent all DNS usage. I can name multiple efforts heading to the direction of circumventing DNS entirely, but I have never been given scientific supporting evidences showing how, and how much, the use of DNS names actually contributed to the Internet centralization.
Yes the DNS namespace has a governance body to ensure DNS names' uniqueness. As far as I know, that namespace governance has not stopped anyone from getting a DNS name; many of my colleagues have got their personal DNS names, so have my students, so have all the research projects I participated in (e.g. https://named-data.net/, Named Data Networking, the most important and exciting project in my whole career that started in 2010 and I am still working on it!)
We are in computer science. Scientific endeavors require supporting evidences to draw conclusions. Therefore I urge everyone, who works on developing solutions to circumvent DNS usage as his/her contribution to Internet decentralization, to first identify quantifiable evidences to demonstrate how, and how much, the use of DNS names (at least before the recent development of DOH, DNS over HTTPS) have actually contributed to the Internet centralization, before you commit your effort.
(DOH, on the other hand, is not part of the DNS design and is a totally different story.)