I notice SFDB blocking IPs from hotmail, gmail & yahoo. Is there anyway I can set in SF for whitelist/skip those big boy IP from SFDB checking? I still want to remain all spam checking except SFDB. Can it be done?

Do SF Whitelist - Excluded Domain/IPs support netmask IPs like 123.123.123.0/16?

 Thanks

 

We too are seeing a huge increase in hotmail, gmail and aol getting on the SFDB.  Perhaps, since there is greater possiblilit of them getting reported due to their volume of message traffic, we could request that the algorithm that logs from those IP's be adjusted somehow.

Our SDFB server actually overrides the minimum "SFDB Network Reliability" setting to ensure it is not too low.

We currently are using a threshold of 9. This means that if users configured a minimum threshold of 3 in their SpamFilter, our SFDB server will increase it to 9 to lower the number of false positives.

If a user decided to be very conservative, and configure a higher threshold of 15 for example, our SFDB server will honor the higher threshold.

This is done because as more and more companies use SpamFilter, more and more are going to report spam, and a higher network reliability is required to obtain more accurate results.

"Unfortunately" the SFDB is very accurate. If an IP is blacklisted, it is **extremely** likely that the IP was used to send spam within the last 24 hours. If the IP belongs to a large provider however, it will cause the issues mentioned here. A mitigating factor is that the large providers often hame several SMTP servers scattered throughout the internet, so when one of their IPs is blacklisted, it only affects a small portions of their customers.

We're listening to your comments, and are brainstorming here to see what can be done about this. The issue is knowing that someone is spamming, but somehow continue to accept emails from them...!

With only 3 reject IDs you'd be seeing many more false positives. This is because we now have many more companies using the SFDB than 6 months ago... Since we now have more reporters, we have been incresing slightly each month the minimum number of rejects necessary to blacklist the IPs.

Say now we have 10 companies running SpamFilter and reporting to the SFDB. A minimum of 2 reject IDs would have probably given "OK" results. Next week the reporters increase to 100. It's very likely now that 2 of those 100 companies can have a false report for that IP, so we had to increase the minimum number or reporters to get less false positives, let's say 5. If the following month we have 1,000 companies reporting, we have to increase it even more, as it's possible that with only 5 companies reporting someone out of 1,000 could yield false positives...