Facebook's isn't quite something straight out of Dally and Towles, but it's the kind of thing you could imagine building with that kind of knowledge. It hasn't been long enough since FB published their topology for people to copy them, but is the idea obvious enough that you'd expect it to be independently "copied"?
Where I can get answers to this stuff1? That's not a rhetorical question! I'm really interested in hearing about other resources!
Alex Gaynor set up a GitHub repo that attempts to answer this entire question. It answers some of the questions, and has answers to some questions it didn't even occur to me to ask, but it's missing answers to the vast majority of these questions.
For high-level answers, here's Tali Garsiel and Paul Irish on how a browser works and Jessica McKellar how the Internet Works. For how a simple OS does things, Xv6 has good explanations. For how Linux works, Gustavo Duarte has a series of explanations hereFor TTYs, this article by Linus Akesson is a nice supplement to Duarte's blog.
One level down from that, James Marshall has a concise explanation of HTTP 1.0 and 1.1, and SANS has an old but readable guide on SSL and TLS. This isn't exactly smooth prose, but this spec for URLs explains in great detail what a URL is.
Going down another level, MS TechNet has an explanation of TCP, which also includes a short explanation of UDP.
One more level down, Kyle Cassidy has a quick primer on Ethernet, Iljitsch van Beijnum has a lengthier explanation with more history, and Matthew J Castelli has an explanation of LAN switches. And then we have DOCSIS and cable modems. This gives a quick sketch of how long haul fiber is set up, but there must be a better explanation out there somewhere. And here's a quick sketch of modern CPUs. For an answer to the keyboard specific questions, Simon Inns explains keypress decoding and why you can't press an arbitrary combination of keys on a keyboard.
Down one more level, this explains how wires work, Richard A. Steenbergen explains fiber, and Pierret explains transistors.
P.S. As an interview question, this is pretty much the antithesis of the tptacek strategy. From what I've seen, my guess is that tptacek-style interviews are much better filters than open ended questions like this.
Thanks to Marek Majkowski, Allison Kaptur, Mindy Preston, Julia Evans, Marie Clemessy, and Gordon P. Hemsley for providing answers and links to resources with answers! Also, thanks to Julia Evans and Sumana Harihareswara for convincing me to turn these questions into a blog post.