[LINK] spam stats

[Ivan Trundle]

On 29/10/2006, at 3:12 PM, Craig Sanders wrote:

> i just checked ran my spam-stats script on last week's mail log file.
> this is on my small home server, which handles mail for about half a
> dozen users (friends and family) in 4 domains.

That seems roughly consistent with my observations: I have a similar  
setup.

>
> in one week, 49225 spams were blocked. that's 168 per hour.  4.9 per
> minute. or about one spam every 12 seconds. a further 94 got  
> through my
> postfix smtp-level filtering to be tagged by spamassassin and diverted
> to a spam-trap folder. about half a dozen got through both postfix and
> spamassasin rules and were delivered to a real mailbox.

I use postfix and dspam, and find that it captures around 97% of all  
spam, across all accounts and clients.

> all this to deliver just under 900 legit messages to half a dozen mail
> users on my system.

My ratios are roughly the same, though I get less real mail (sigh)  
overall.

But what it does is provide a graphic picture of the amount of  
unwanted traffic on our networks. It's bad enough that here in  
Australia we have to have fraudband, but the extra overhead is  
unsustainable, and must surely be affecting our overall network  
capacity.

> i estimate that i spend at least 4 or 5 hours per week examining my
> spam-traps, creating and updating anti-spam rules, monitoring log  
> files
> - all so that email remains usable, and not buried under the deluge of
> spam. i've been doing this for over 10 years, sometimes more than 5
> hours, sometimes less...but it's a constant drain on my time and  
> energy.

Ditto, though probably less time per week - closer to 2 hours. The  
corpus is well-tuned, and I spend less time now checking for false  
positives (or maybe I don't care any more!)

> this is outrageous. i'm sick of it. unfortunately, i've got no choice
> but to continue doing it because i don't want to give up email. and
> there's no other way to do it, anti-spam laws are basically  
> ineffectual
> and always will be - even with spammers like Mansfield in Perth being
> fined millions, he isn't and wasn't even a significant part of the
> problem. he was a big spammer here in australia, but insignificant
> compared to some of the american spammers.

I concur. And don't forget Chinese spammers - which I spend most of  
my time dealing with (having worked with a Chinese colleague).

It is absolutely appalling, a waste of time, and a waste of  
bandwidth. Imagine the outrage if junk mail in letterboxes amounted  
to the same quantum.

As far having to trawl through the detritus of spam, I avoid it like  
the plague - and it sickens me that there is a market for all this crap.

On the bright side, at least most spam is text or image-based, and  
not moving images (it will come, if it hasn't already).

But again, imagine if any other infrastructure was equally clogged:  
roads, rail, air, water, etc. It's madness.

As an interesting aside, I have made some observations about the TYPE  
of spam that each of my server accounts receive - this IS  
illuminating. My teenage daughter, for example, gets mostly slimming  
spam (which she certainly doesn't need - but it seems that there is  
such a thing as teenage spam), A few friends get mostly Israeli  
investment scams and viagra/sex, I get viagra and Nigerian spam,  
others get a combination of viagra and sex organ improvements, whilst  
a few others get mostly viral content. If I hadn't checked, I would  
have assumed that the type of spam would have been equally spread  
across all of the accounts.

There is, naturally, a direct correlation between the age and  
exposure of the e-mail address and the level of spam. Having a public  
e-mail address (as I have done for many years, being a web manager  
etc) and activity in e-lists (such as Link) increases the exposure to  
spam, too. All these things are obvious, and not easily avoided.

But what annoys me is that whilst some e-mail addresses are  
disposable, and can be left behind, some e-mail addresses tend to  
stick for life - I'm wondering if it is wise to have an address that  
uniquely identifies you for your lifetime (such as john at howard.com,  
or tony at barry.id.au, etc).

Food for thought.

Warmly

iT