Computer Networks: A Systems Approach (2019) | Hacker News Reader

Computer Networks: A Systems Approach (2019)

book.systemsapproach.org

128 pointsskilled10 comments

Open articleView on HN

The book is really cool, but somewhat surprisingly it doesn't even mention NAT.

the textbook is better, I think that might mention it but I don't recall

https://book.systemsapproach.org/internetworking/basic-ip.ht...

But ARP and NAT are not the same

Mentioned in passing in 8.5.6. Firewalls

https://book.systemsapproach.org/security/systems.html#firew...

Co-author of the book here. Thanks for the comment. I agree, it does seem surprising to miss NAT. I opened an issue (which we will address at some point): https://github.com/SystemsApproach/book/issues/75

Traditional textbooks on networking are so verbose and tedious... often very focused on theory and less helpful for every-day networking.

I would love to see courses based on tools like shadow [1], encouraging a more hands-on learning experience. Network simulators seem pretty cool, but it is hard to benefit from them using the docs alone. With shadow it seems I would be able to run actual software on a simulated network.

1: https://github.com/shadow/shadow?tab=readme-ov-file#why-is-s...

My group used to teach Computer Networks at a university ([1] gives a fair overview of the scope if you know Croatian or use a bit of Google Translate). We used CORE and PHP built-in web server [2] for the basic course. IMHO, stuff like Shadow is too overwhelming for the introductory course, but we planned to use it together with ns-3 and SimGrid for the advanced course (which in the end we never got a chance to develop for unrelated reasons).

[1] https://group.miletic.net/hr/nastava/kolegiji/RM/

[2] https://coreemu.github.io/core/

Computer Networks: A Systems Approach - https://news.ycombinator.com/item?id=26110796 - Feb 2021 (23 comments)

The Book “Computer Networks: A Systems Approach” is now open source - https://news.ycombinator.com/item?id=17824472 - Aug 2018 (14 comments)

For something that's supposed to propose a "novel" approach, this sure does stick to the arbitrary definition that network protocols are restricted to time sensitive protocols.

The application layer shouldn't be an after though or a miscellaneous category to which we give almost no attention. The application layer, along with the user, should drive the network.

Reply on news.ycombinator.com