North American Network Operators Group|
Date Prev | Date Next | Date Index | Thread Index | Author Index | Historical
Re: SMTP store and forward requires DSN for integrity (was Re:Cluelessanti-virus )
Douglas Otis wrote:
Near enough so that reject at SMTP data phase will cover all legitimate cases that may exist.On Dec 9, 2005, at 10:15 AM, Todd Vierling wrote:1. Virus "warnings" to forged addresses are UBE, by definition.This definition would be making at least two of the following assumptions: 1) Malware detection has a 0% false positive.
2) Lack of DSN for email falsely detected containing malware is okay.
1 dropped email is NOT worth thousands of to-forged addresses DSN's The admins that care will design their systems to reject by smtp DATA pohase
3) Purported malware should be assumed to use a forged return-path.
Yup, thats true of everything in the wild today.
Exactly, only then is DSN's acceptable to otherwise near guaranteed incorrect destinations.4) The return-path can be validated prior to accepting a message.
Obeying the SMTP standard should not generate crap unneedlessly.5) SMTP should appear to be point-to-point.
That means reject virus at data phase, discard them afterwards since the DSN violated the much more important rule not spewing crap at innocent 3rd parties.
To generalize it:6) MTAs with AV filters are the only problem.
Systems which generate DSN's to known incorrect destinations are the EXACT problem discussed here. Currently that fits all "scan after receipt of messafe" av systems that send DSN's
I havent been keeping on top of the sender/return path verification scene.
But blacklisting abusive systems to create pressure on admins to turn the spewage off is the current time honored mechanism.
If you can't trust AV handling of DSNs, why trust their detections?
One is fairly easy to measure. The other SHOULD be fairly easy to turn off.
I would rather my MTA return the DSN it generates when it receives your MTA's smtp rejection to the data command contents.Would you rather see emails simply disappear?
2. It is the responsibility of the *SENDER* not to send UBE.
Blocklisting them should even things out eventually.