Re: [Resolverless-dns] Paper on Resolver-less DNS

[Ben Schwartz <[e-mail]>]

Thanks for conducting this investigation, Erik!

In my view, the two main concerns with resolverless architectures have been
(1) simplifying stolen key attacks and (2) potential interference with
DNS-based load balancing.

Currently, an attacker who has stolen a TLS private key also needs control
of the victim's network in order to mount an impersonation attack.
Resolverless DNS (especially without DNSSEC, as in your model) might make
these attacks much easier to execute, by sending the victim a DNS record
that points the compromised domain to the attacker's IP address.

Domains that rely on DNS-based load balancing have concerns about third
parties handing out DNS records for their domains.  These third parties
might (accidentally) provide the same DNS record to all of their users,
overloading a particular server instead of spreading the load across many
servers.

If you're interested in this topic, I hope you'll think a bit about these
problems, and see if you find a way to solve them.

On Tue, Aug 13, 2019 at 8:07 AM Erik Sy <[e-mail]>
wrote:

> Hi all,
>
> I wrote a paper to clarify the design of resolver-less DNS and to
> investigate its feasibility.
> You can find the paper here:
>
> https://svs.informatik.uni-hamburg.de/publications/2019/2019-08-13-Sy-preprint-Enhanced_Performance_and_Privacy_via_Resolver-Less_DNS.pdf
>
> In total, I think this is a great performance-optimization especially
> for clients on high-latency access networks. Furthermore, it can improve
> the client's privacy posture towards the recursive DNS resolver.
>
> Best regards,
> Erik
>
>
> --
> Resolverless-dns mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/resolverless-dns
>