Instead of myriad of linodes all trying to block spam in parallel, why doesn't known spam get blocked at the data centers before hitting the individual linode instances?

After all, if the block lists are 99% effective, why are we all wasting resources when the data centers could be doing that for us? Seems like a terrific savings to all the linodes.

The linodes would still run the same, but most of the pointless wasted traffic would be gone.

what defines spam though? different customers might want to do different things: you might just drop it, i might absorb it to do analysis on, etc.

IMO data centers shouldn't be messing with data, just passing it.

I'd think the sheer number of rules needed would also put a heavy load on the core routers (or host machines, if implemented there), too,

My sentiments exactly.

_________________
/ Peter

I don't think there is such thing as "known span".
Actually, I may very well close my accounts if such scheme were implemented lest I would miss some so-called "known spam" messages.

I agree. One of the main reasons I started my linode several years ago was I got sick and tired of ISPs messing with my email. I'll make my own decisions, thank you very much.

Agreed, I'll do it myself. My current spam-detection is far and away better than anything I've seen from an ISP, because I can custom train the bayesian filter rules and set my whitelist and not have to rely so much on blacklists, which is a good thing since they can result in a lot of false positives.

_________________
----
Ed/Bones.