Re: New proposal for the interactive message tracking protocol
Alexey Melnikov <[email protected]> Mon, 14 Feb 2000 13:21:37 +0000
| Newsgroups | gmane.ietf.msgtrk |
|---|---|
| Message-ID | <[email protected]> |
Lyndon Nerenberg wrote: > > >> 2. The tracking activity is kept away from the e-mail > >> system, so that an SMTP server cannot be saturated with > >> tracking requests by malicious users. > > But since tracking the message is intimately tied to SMTP AUTH, > I don't see how this could be done (easily) outside of SMTP while > staying inside the framework of SMTP AUTH. That is right. Any protocol that will be used/invented for message tracking must have authentication. Seems like a requirement. > >> > 3. It requires only slight changes in SMTP, namely to > >> generate the logs. > > Alexey> In most cases MTA already generate logs. > > However you cannot assume the MTA generates logs anyplace accessable > to you. I could very easily ship all the logging info to a seperate > write-only log server that the tracking software has no access to. This is a good point, but it doesn't really prove anything. For example, the protocol used for processing tracking requests may run on the same computer where syslog data is stored. > And for popular MTAs that log through syslog(3) (e.g. sendmail), > you cannot trust the log data unless you've secured your entire > syslog environment. There was "secure syslog" BOF in Washington IETF. So sooner ot later the problem will be solved. Alexey