Hi Helge,

for calls is quiet difficult to make a spam filtering since you do not have enough information to feed the filters. IMHO the only available data is the source IP and the FROM header - you can use permissions module to set filters. another idea is to credit a call with some trust based on agreements with peers. You can read more on this concept in the TLS documentation - you can extrapolate it from TLS to IP based auth.

for MESSAGEs...different story as is very similar to email - you may try running a spam engine (as spamasassin) via external commands (exec) - but this an highly inefficient way to hook it....but if you manage to do it - at least as a proof of concept - we can think to more efficient ways ;).

best regards, Bogdan

Helge Waastad wrote:

Today, just some refelctions...

A colleague of mine, early today, started getting spam messages on his Skype account, and this got me thinking:

What about spamfiltering on proxy level?

Has anyone tried to execute a spamassassin script on, for instance, a MESSAGE and either forwarded or discarding the call?

I know this would have a deep impact on processing time, but anyway, I thought this was an interesting scenario..

br hw