Problems sending mail to yahoo?

Steve Atkins steve at blighty.com
Mon Apr 14 00:27:05 UTC 2008



On Apr 13, 2008, at 5:04 PM, Barry Shein wrote:
>
>
> Massive quoting gets old fast so I'll try to summarize and if I
> misrepresent your POV in any way my profuse apologies in advance.
>
> First and foremost let me say that if we had a vote here tomorrow on
> the spam problem I suspect you'd win but that's because most people,
> even (especially) people who believe themselves to be technically
> knowledgeable, hold a lot of misconceptions about spam. So much for
> democracy.
>
> I say the core problem in spam are the botnets capable of delivering
> on the order of 100 billion msgs/day.
>
> You say there are other kinds of spammers.
>
> I'll agree but if we got rid of or incapacitated the massive botnets
> that would be a trickle, manageable, and hardly be worth fussing
> about, particularly on an operational list.

> The reason is that without the botnets the spammers don't have address
> mobility. You could just block their servers.

Address mobility doesn't buy you that much. It's relatively easy to  
mechanically
detect, and block, IP addresses that source mail solely from spam- 
related
botnets. (Not easy in the absolute sense, but easier than other problems
and, mostly, a solved one). Botnet sourced mail generally doesn't get
seen much by recipients at ISPs with competent spam filtering. It sure  
can
cause other operational problems, but in terms of being a "spam problem"
it's not the biggest one out there.

Blocking unwanted mail from sources that send a mixture of wanted
and unwanted mail, while still allowing the wanted mail through is
extremely difficult, and a much, much harder problem for spam
mitigation to solve. And those are primarily the non-botnet sources.

Spam filtering at real ISPs with real recipients has to deal with the
fact that recipients do want to read some of the mail they're sent
from Gmail, Yahoo Groups, Topica and suchlike.

Cheers,
   Steve






More information about the NANOG mailing list