[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: [OT] Yahoo's Antispam proposal



s. keeling wrote:

Even at %65 (according to Economist/Brightmail) of overall traffic, spam
is still very manageable with the right software.

though, or my mailbox will overflow.  When it reaches 99+% I shall have to
give up email entirely.


Worst case, you can pay somebody to do it for you, one way or another.
But abandon email because of spam?  Not from my vantage point.



That's the key point to all of this.

Spam is a Billion Dollar industry.
Spam-Blocking is also becoming a Million Dollar industry.

If you don't want spam, you have to pay someone.
It's the "pay" part that gives a lot of heartburn.

And you comment about 65% being manageable. Maybe for you. But if you are trying to process email for 1,000+ accounts this quickly adds up to major hardware and network investments just to manage the cruft. Without the cruft it would be easy.

Example: I use amavis + spamassassin + clamav for one of my filtering steps. This has a hard limit of 4 email requests at a time and ~30 seconds per email. This puts me at a definite limit of 86400/30*4 11,520 emails per 24 hour period which means that my current system would only allow 1,000 email account to send/receive ~12 emails per day.

Without filtering of this nature and using say only bogofilter in procmail scripts, I can manage 20 emails per second or 1.7 million emails per day. That's a 1000X increase in performance. I use a Via Epia 533 fanless box for my mail server and it works wonderfully for my home account. But that's only as a point of reference on my hardware.

To get around this, I would need to invest in a box that is going to cost me 20X what this box cost me. So now your trivial spam load has cost me several thousand dollars in hardware and if I really did have to manage 1,000+ email accounts I would have to pay extra for the excess bandwidth usage of the "cruft".

Spam is Very Expensive.



Reply to: