On Mon, 25 Mar 2024 15:58:33 -0700, Gerald Oskoboiny via mailop <mailop@mailop.org> wrote:
>Is it still necessary to warm up new IP addresses gradually >instead of going directly to this volume of deliveries? My >impression is that it's less and less necessary in the age of >DMARC, SPF and DKIM. The rule that governs many of the dynamic IP reputation systems that I am familiar with is: "Don't give us any surprises." When I was a spam analyst at MSFT in a previous geologic era, we had a guy who would build out a new Eonix /24, send test messages to seed accounts until he decided that he had found ways around the rules that killed the tail end of yesterday's blast, made his changes, backed his truck up to our network and dumped between five and fifteen million messages over a period of about half to three quarters of an hour. At that time, we had nothing technical implemented that would handle this, so it worked quite well. Eventually, we were able to convince people who did the engineering at the border to consider the "No Surprises" rule. One of my clients, without consulting aforehand, apparently decided that he really needed to do a 10X augmentation to his daily volume. Before the inevitable algorithmic corrections based on the ghastly volume of spam notifications, the border logic at several major providers moved his IP reputations from Good or OK to reject, with sampling. Overall, his border rejection rate went from 1.45% (not great, but not yet a policy enforcement matter) to 55.6% (yes, this is a policy enforcement matter). The sudden onslaught you propose may actually succeed in the main, and after a couple weeks of zero-complaint/excellent-open stats you will be back in good graces overall, it might be well to look at a week-long cutover transition, if the technology permits. mdr -- Ad finem pugnabo. _______________________________________________ mailop mailing list mailop@mailop.org https://list.mailop.org/listinfo/mailop
