Hacker News new | past | comments | ask | show | jobs | submit login
The Twelve Networking Truths (ietf.org)
159 points by moks on Nov 2, 2015 | hide | past | favorite | 27 comments



A good complement to L. Peter Deutsch's of Sun Labs Eight Fallacies of Distributed Computing:

1. The network is reliable

2. Latency is zero

3. Bandwidth is infinite

4. The network is secure

5. Topology doesn't change

6. There is one administrator

7. Transport cost is zero

8. The network is homogeneous

He noted "All prove to be false in the long run and all cause big trouble and painful learning experiences."


"In protocol design, perfection has been reached not when there is nothing left to add, but when there is nothing left to take away." This seems like a good chance to acknowledge the author, Antoine de Saint Exupéry.


St-Exupéry never mentioned "protocol design". This is a pastiche, not a plagiary.

As to why St-Exupéry's name was not included, I suppose the editor thought the reference was obvious, which it is indeed.


Anyone in technology should read the whole chapter 3 of Wind, Sand and Stars, if they haven’t before:

http://pastie.org/pastes/10477072/text


Number 5 strikes home in today's world of big monolithic programs which each aim to solve all the world's problems...

"(5) It is always possible to aglutenate multiple separate problems into a single complex interdependent solution. In most cases this is a bad idea."


(11) Every old idea will be proposed again with a different name and a different presentation, regardless of whether it works.

I see this all the time. I'm pretty sure there have been no really original ideas in computing since the 1970/80s.


Distributed hash tables are a recent idea in networking

https://en.wikipedia.org/wiki/Distributed_hash_table#History


My dad mentioned that about automotive design, pretty much everything interesting had been invented by 1940. After that it's all new manufacturing technologies, materials, or economies of scale. Or some other technology from another field. (Good example is developments in power electronics meant electric cars were viable again after an 80 year hiatus)

Also I like that 'It Has To Work' is numero uno.


Everything that's new is old already.



There should be errata to the errata since the OPERA result of neutrinos traveling faster than c has since been refuted [1].

[1]: https://en.wikipedia.org/wiki/Faster-than-light_neutrino_ano...


What's the deal with #3 about flying pigs? I mean, I see why it's true (ish), but I don't see the specific relationship to network protocols.


I think it means that you can take a bad foundation and push it forward amazingly far with enough effort, but you're setting yourself up for lots of problems if you do so, and you're usually better off starting from a solid foundation.

For a concrete example related to networking, transmitting bits over telephone lines designed for voice can get you amazingly far, but it still sucks a lot and you'll eventually hit a solid wall.


Given this statement:

These truths apply to networking in general, and are not limited to TCP/IP, the Internet, or any other subset of the networking community.

Its not hard to come up with many examples particularly bloated one (ie pig like). For example... CORBA and SOAP.


Well, it is an April 1st RFC, or perhaps he could be making the point that you can do things like make telnet work over UDP but why would you?


OP, please append "(1996)" to the title.


Universal Truth of the ISP:

- 97% of the time the issue is on the customer network

- its not the network, ex. you cant transfer files at 10g because your using single 7200 RPM HDD on each end :[

- often, the best network people in the company will be the farthest from customer interaction. Customers repel them with almost magnetic physical force.


I feel like these truths expand to many things, far more than just networking alone. Especially 2a and 11.


Oh boy, this is 20 years old. Why is this even coming up.


Something that is true after 20 years and still noteworthy seems like a good read to me. I agree the OP should add the year though.


rule (11) Every old idea will be proposed again


Santayana explains.


a universal truth rarely ceases to be universal or true.


The language used in the introduction alone was enough to make me check the date of publication... Those IETF jokers!


This document was incorrectly filed as an April 1 RFC, when in fact it should be an informational RFC.


"the Internet community".


Can someone please explain it to me?




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: