Understanding email routing (i.e. MX-records)

wireload.net

17 points·mvip··7 comments

Open articleView on HN

It's a lot more fun that that, a single hop is barely 'routing' :)

For bonus points

Go back in history to when mail was literally routed hop by hop through SMTP? How'd that work?

What is the sending SMTP client expected to do when the server is unreachable?

Have you seen the headers of a mail your receive on gmail? How many smtp hops inside gmails network?

In an enterprise, how many smtp servers does your email hit before it exits?

And that's partly how the horror of sendmail evolved :) Not even going into handling multiple email protocols and incompatible encoding types.

Would've been nice with a little more meat .. like convering all the steps performed when detecting the actual SMTP servers (including connection attempts, alternative lookups when no mx records exist etc).

While that may be a fun read, it's hardly relevant for the average user (or even admin). If you're at that level, I suggest reading relevant the RFCs. :)

If I write an article called "Understanding how websites work" where I just regurgitate the definition of an A record can I be on the frontpage of HN too?

Yes. Part 2 could cover CNAMEs.

If you receive a sufficient number of up-votes, yes. I would recommend using simple language, and avoid digressing too deeply in to networking concepts. Keep in mind that HN is full of hackers, but it's also full of product people, sales people, and new-comers.

This is a good high-level write up for people who might know a little about how DNS works because they've configured a domain, but don't know a lot of detail about the different record types. I can see how a lot of people would find it interesting.

You don't even need to address common problems or have any remotely interesting commentary whatsoever!

Reply on news.ycombinator.com