Hacker News new | past | comments | ask | show | jobs | submit login
One man's spam - graphing spam rates since 2002 (solipsys.co.uk)
18 points by RiderOfGiraffes on Jan 30, 2011 | hide | past | favorite | 10 comments



My mail server does primary MX for a couple of small domains (5 or so people) and relays a few domains for friends to gmail.

I've been doing some crude graphing for the last couple of years based on specific text entries in Exim's log:

http://cmsj.net/stats/mail.png

Note: I moved the mail delivery to a new machine about 3 weeks ago which might have upset things, but the spam levels had collapsed before that. I'm somewhat tempted to filter out the peak level because it hides so much nuance in the rest of the graph.


hit post too early...other notes: - yes, I do sender verification. no, I am not prepared to debate it with you, live with it ;) - I've always been a bit surprised that there aren't more malware rejections. Either ClamAV sucks or people aren't trying very hard to give me viruses. - I find it interesting that the 'blunderbus' approach of sending to all imaginable addresses (i.e. the blue section of the graph) has diminished significantly over time.


  > hit post too early ...
You know you can edit for up to 2 hours ...


I did not know that! Unfortunately I missed it this time, but I will do that next time :)


Oddly, that's what the "edit" link means next to your item or reply when you've first submitted it. Lots of people miss it, not least because people often don't check what they've written.

<fx: grin>


The horizontal axis on that graph really needs to be clarified. Yes, I can figure out what the 1s represent, and count the years, but...


Agreed - I was dealing with the data and thought - "Hmm - HN might be interested in a graph of this" so I knocked one out quickly. The scales don't really matter. It covers from 2002 to 2010, and really there's not much else you need to know.

In case you want, the raw data are here:

http://www.solipsys.co.uk/new/RawSpamCount.html?HN

The first field is the number of emails, the last two are year and month. The others are meaningless.


You got 50 000 spam messages in one month? Is this just for one address or multiple?


I have a complicated setup, but in essence, I have, by necessity, ended up with a catch-all on a domain. There is value to what I do, but the cost is the level of spam. It's like having hundreds of honey-pots.

I'm considering washing them of personal information and making them available.


I used to get insane amounts of spam, but seeing this graph made me look again. Based on procmail logs, it looks like about 29,000 pieces last month, which is definitely down from before.

About 90% of my incoming spam in on two email addresses.




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

Search: