Currently we have hundreds of domains from hundreds of ISPs sending
email from and to and through our servers. How do we make a global
change to SPF without shutting everybody off in a massive email outage.
Also we want to use SPF to decrease the load on the back end servers
that have to handle unwanted email, but our front-end mail scanners are
already nearly CPU-Bound. There has to be a way to reduce the amount of
spam that reaches the SPF servers to compensate for the extra processing
and latency that SPF would cause. Does anybody have some actual numbers
that would allow us to say SPF is going to decrease spam by a certain
percent and it will take a certain amount more hardware or so many
percentage decrease of incoming spam to compensate?
Also, from what I understand SPF isn't going to decrease at all the
number of spams that our mail scanners have to process. It only
validates that the from address is valid for certain domains (i.e. We
will know when an email is being spoofed from a domain that uses SPF.
However, for the larger percentage of incoming email which originates
from fake or non-SPF domain names: what process is used to flag those
first?)
What does everybody else use besides SPF to decrease the amount of spam?
If you want to derail as a CTO, just make the CEO feel like a ninny.
-- Ted "Dililies" Sept 6, 2004 InfoWorld, pg 12