mail-vet-discuss
[Top] [All Lists]

Re: [mail-vet-discuss] Seeking consensus on MUA use

2008-12-17 21:25:42
Murray,

While not including local-parts will help ensure the Authentication- 
Results header itself is less deceptive, this does not prevent use of  
Authentication-Results header to annotate PRA or the MAIL FROM email- 
addresses.  Unfortunately, annotations applied by MUAs will encourage  
confidence artist to gain access to the authorized SMTP clients.  Weak  
security will result in systems being compromised on an ongoing basis,  
as is currently the case.  When providers are contacted regarding  
security issues, their response might be slow or never seen.

When access to an SMTP client has become compromised, there is no  
Reverse Authorization Directory that lists the potentially thousands  
of domains being put at risk.  Even if there were such a list, the  
time needed to discover and contact the affected domains, and for them  
to then inform theirs DNS providers of required changes, will be  
further delayed by TTLs of SPF records.  When dealing with fast moving  
bad actors who start and finish their deeds within brief periods, the  
speed in which reputation information is conveyed represents a vital  
aspect of any effective mitigation effort. There are IP address  
reputation services that are able to respond within a few hundred  
seconds of a problem being detected through automation.  This would be  
well ahead of most MUAs.

A response based upon a domain name can not be done within a timely  
manner, if at all.  Blocking at the domain would be highly disruptive,  
and would prevent further communications that might indicate the  
nature of the new threat.  The only salient, timely, and effective  
identifier which can establish a transient negative result regarding  
trust would be the IP address of the SMTP client.  This would be the  
IP address authorized by either RFC 4408 section 2.4 (SPF),  or RFC  
4406 section 4 (Sender-ID).  To ensure the reputation of the SMTP  
client IP address can be checked, the SMTP client IP address  
authorized by RFC 4408 or RFC 4406 MUST BE captured by the  
Authentication-Results header.  Nothing else would ensure a means to  
obtain the information needed to assess either SPF or Sender-ID  
results.   This is essential to assess the reputation of the  
authenticated IP address of the SMTP client.  Nothing else other than  
the IP address of the SMTP client will have been authenticated within  
SPF or Sender-ID results.

Why should an Authentication-Results header cater to providers who  
hope to profit from marketing an allusion of security, rather than  
establishing a means for security?  Adding the IP address of the SMTP  
client allows reputation services a means to indicate which method  
appears supported by each SMTP client.  In addition to dealing with  
security breaches, the reputation service might indicate that an an  
SMTP client does not appear to limit use of the Purported Responsible  
Address, or the MAIL FROM.  These aspects of access and use will play  
a significant role in what is safe to annotate, and perhaps how it  
gets annotated.

The IP address of the SMTP client represents the _basis_ used for  
upstream evaluations of Sender-ID and SPF.  Including the IP address  
is essential when evaluating a source expected to impose restrictions  
upon either the PRA or the MAIL FROM.   The IP address is essential,  
because not all SMTP clients will impose the same restrictions, and  
where access to any SMTP client might become compromised.  Treating  
SMTP clients only by the authorizing domains is not practical and can  
not deal properly with the realities of inconsistent protocol adoption  
and fragile security.

The IP address used by either Sender-ID or SPF has been weakly  
authenticated by way of the SMTP exchange itself.  For either SPF or  
Sender-ID, only the IP address of the SMTP client will have been  
authenticated.  Whether the PRA or MAIL FROM even offered an  
authorization is based upon a presumption of the scope being applied,  
and the access restrictions imposed.  Section 4.1 of the  
Authentication-Results Header Field draft states:
,---
An MUA SHOULD NOT reveal these results to end users unless the results  
are accompanied by, at a minimum, some associated reputation data  
about the authenticated origin identifiers within the message.
'---
For either Sender-ID or SPF, the only authenticated origin identifier  
is the SMTP client IP address!!!

Identifying compromised systems or protocol implementations must be  
dependent upon the SMTP client IP address, and not that of some  
authorizing domain.  In other words, restrictions imposed upon MAIL  
FROM or the PRA are dependent upon the authenticated origin  
identifiers within the message, the SMTP client IP address.  When  
adding either an SPF or Sender-ID result, it is of paramount  
importance for vetting the origination of the message, that the IP  
address of the SMTP client be included.

Changing the Authentication-Results header to include the basis of the  
SPF or Sender-ID results would help ameliorate potential harm of  
uncertain scopes, assumed authentication, or when coping with security  
breaches.  This concern is not about dead horses,  or of information  
unrelated to the use of SPF or Sender-ID.   The SMTP client IP address  
information represents the basis for Sender-ID and SPF   
authorization.  Ensuring the SMTP client IP address availability is  
essential to meet the requirements set forth in section 4.1 of the  
Authentication-Results Header Field draft.

Not listing this information borders on negligence.

-Doug



_______________________________________________
NOTE WELL: This list operates according to 
http://mipassoc.org/dkim/ietf-list-rules.html 

<Prev in Thread] Current Thread [Next in Thread>