I went through this a few years ago, deciding that I had only a piecemeal understanding of DNS based on the specific things I had run into. I knew about `dig(1

I went through this a few years ago, deciding that I had only a piecemeal understanding of DNS based on the specific things I had run into. I knew about `dig(1)` and BIND and a CS101 idea of how recursive DNS resolution works, etc. But I was missing the working knowledge needed for designing and implementing anything non-trivial or for debugging non-working systems. So I read "DNS and BIND" (not quite cover-to-cover, but close, skipping over the details on some of the more exotic features). I set up a real BIND server for some unimportant personal web sites. None of it was hard, but it did involve a bunch of (time) investment. To be clear, BIND is not the right thing for many use cases, but a lot of DNS ideas and terminology still come from BIND and I found that stuff very valuable. I think books are underappreciated for learning stuff like this. Most resources you find on the web are high-level theory ("here's a block diagram of how recursive lookup works"), task-oriented ("how do I get `dig` to do a recursive lookup"), or otherwise low-level (e.g., reading the source of your local DNS client to understand its retry policy). To understand the pieces and how they fit together, from the theory (what they're trying to achieve) down to implementation (what caches exist where), I find there's no substitute for a holistic approach that you usually find in books, but rarely (not never) find on the web.