Not necessarily. For example, if I access 142.250.187.228 what domain name was in the Host header. Hint: It's not www.google.com. If I visit 65.9.141.10, then what domain was in the Host header. Hint: It's not search.brave.com.
But do not take a random HN commenter's word for it. Try this yourself. Try to map IP addresses to domain names on a very large sample. I have tested this on hundreds of thousands of domains. It will not work reliably. Not even close. Compared to SNI, which is both simple and reliable, it is a total PITA.
If this 1-to-1 IP to domain mapping and 100% reliable reverse DNS idea made any sense, then why claim encrypted DNS offers "DNS privacy". Anyone can see IP addresses on the wire. If all IP addresses can easily be mapped to domain names, then how would this provide any privacy.
Mapping IP addresses to domain names has to account for the fact that IP addresses can change. To make sure the mapping is correct one has to constantly keep doing lookups. SNI stays the same even if the IP changes. It requires very little effort; it makes the process of determining the domain name trivial.
The goal is not total privacy. (Impossible when using the internet anyway.) The goal is to not to make surveillance easy. That's what SNI does. It makes surveillance (and censorship) super easy.
If I am wrong about this, and SNI is AOK, then why was ESNi developed and why is ECH being developed. It is not an easy problem to solve. Why bother. SNI was a feature to benefit websites, not website users. It has some very undesirable properties for users.
Even with CDNs that need to use SNI, it is still possible to avoid sending the domain name in plaintext on the wire. One can use a dummy name, i.e., not the name in the Host header, as the SNI and the CDN can still return the correct page because it uses only the Host header. This was nicknamed "domain fronting". I was using this technique before anyone started calling it that. Using a dummy name can reduce the surveillance value of SNI, sometimes to zero. After some "security researchers" started bragging about their discovery of "domain fronting", AWS and other CDNs started checking SNI against Host header to make sure they are the same. Now "domain fronting" no longer works. Countries that censor the internet, who are valued AWS customers, rely on accurate SNI in order to carry out censorship. Go figure.
But do not take a random HN commenter's word for it. Try this yourself. Try to map IP addresses to domain names on a very large sample. I have tested this on hundreds of thousands of domains. It will not work reliably. Not even close. Compared to SNI, which is both simple and reliable, it is a total PITA.
If this 1-to-1 IP to domain mapping and 100% reliable reverse DNS idea made any sense, then why claim encrypted DNS offers "DNS privacy". Anyone can see IP addresses on the wire. If all IP addresses can easily be mapped to domain names, then how would this provide any privacy.
Mapping IP addresses to domain names has to account for the fact that IP addresses can change. To make sure the mapping is correct one has to constantly keep doing lookups. SNI stays the same even if the IP changes. It requires very little effort; it makes the process of determining the domain name trivial.
The goal is not total privacy. (Impossible when using the internet anyway.) The goal is to not to make surveillance easy. That's what SNI does. It makes surveillance (and censorship) super easy.
If I am wrong about this, and SNI is AOK, then why was ESNi developed and why is ECH being developed. It is not an easy problem to solve. Why bother. SNI was a feature to benefit websites, not website users. It has some very undesirable properties for users.
Even with CDNs that need to use SNI, it is still possible to avoid sending the domain name in plaintext on the wire. One can use a dummy name, i.e., not the name in the Host header, as the SNI and the CDN can still return the correct page because it uses only the Host header. This was nicknamed "domain fronting". I was using this technique before anyone started calling it that. Using a dummy name can reduce the surveillance value of SNI, sometimes to zero. After some "security researchers" started bragging about their discovery of "domain fronting", AWS and other CDNs started checking SNI against Host header to make sure they are the same. Now "domain fronting" no longer works. Countries that censor the internet, who are valued AWS customers, rely on accurate SNI in order to carry out censorship. Go figure.