--On Monday, April 30, 2001 8:37 AM -0700 Chuq Von Rospach
<[EMAIL PROTECTED]> wrote:
> On 4/30/01 8:24 AM, "Tom Neff" <[EMAIL PROTECTED]> wrote:
>> From an ISP standpoint, what really matters about spam is the number of
>> recipients. The really noticeable difference between a legitimate
>> mailing list and a spam campaign is that the latter is HUGE.
>
> That's not a safe assumption, Tom. There are huge, legitimate mailing
> lists as well.
But wait, what does "huge" mean? How big are you thinking? The biggest
listserv I know of is TESL-L which has about 12,000 members. If half of
those were on AOL (unlikely given the specialized educational topic, but
possible for a similar sized list) then there'd be 6,000 deliveries
attempted to AOL. 6,000 is POCKET CHANGE for a spammer. A typical
campaign tries to hit ten times that many or more.
>> As a matter of
>> practice, if you check your sender's relay and you impose a limit (hard,
>> agreed-upon or otherwise) on the number of recipients within your domain,
>> you can let real lists through while blocking spam campaigns, or so it
>> would seem to me.
>
> That doesn't scale to AOL, I don't think. I wish I could toss some numbers
> here to illuminate what I'm doing, but I can't. Just realize there are
> really big lists (and then add a couple of zeros) have have really big AOL
> subscriber bases...
The question is whether there are really big legit lists that rival a spam
campaign in size. My guess is no. But I agree size alone won't do it, you
need some kind of authentication. As for scaling to AOL, remember their
hardware and software has to scale in the first place or they don't run!
The measurement of interest would be CPU seconds per delivery, and I
suspect they do rather well compared to smaller gateways.
