That being said, depending on the RBL in-question they can go beyond the
/24 if it's nested within a larger prefix. So you may end up doing work
that doesn't end up making any/much difference to the end result
(if/when listed).
You could split outbound/SMTP to use delivery/outbound address space
that can be kept away from inbound SMTP. That way you can react to
listing/s with (relative) ease, though it tends to pay dividends if
you've "warmed up" the spare delivery IPs beforehand (even with a very
small %age of outbound mail).
Cheers,
Luke Thompson
Operations Manager
On 16/10/21 11:40 am, Grant Taylor via mailop wrote:
On 10/15/21 5:37 PM, Dave Crocker via mailop wrote:
Yes, but... That's the point that is intuitively reasonable which
doesn't make real sense to me, after thinking about it.
What doesn't make real sense to you? The relation of reputation and
spreading services across multiple IPs?
I'm not seeing how 'reputation' interacts for these two functions.
The motivation for spreading service IPs across different /24 prefixes
is so that if (when) enough individual /32 IPs get enough negative
reputation to pull down the /24 aggregate prefix, only the services in
said /24 aggregate prefix would be harmed.
I guess there is some room for asking what is the impact to a service
stack if one (or more) of, but not all of, the services in the stack
are adversely effected. How much degradation can the overall stack
sustain before ceasing to function.
It might make more sense if you had multiple sets of the service stack
and each component was rotated.
E.g.
A1, B2, C3 in prefix 1
A2, B3, C1 in prefix 2
A3, B1, C2 in prefix 3
So the loss of any prefix do to reputation wouldn't take out any
entire stack.
_______________________________________________
mailop mailing list
mailop@mailop.org
https://list.mailop.org/listinfo/mailop
_______________________________________________
mailop mailing list
mailop@mailop.org
https://list.mailop.org/listinfo/mailop
