Hi,

 

Chris asked me as well to review draft-ietf-syslog-dtls.  I agree with
the overall sentiment that this draft is in good shape.  Here are my
comments:

 

 

1.Editorial:  "Syslog" or "SYSLOG" - should use consistent
(non-)capitalization throughout.  

 

2. Introduction, last paragraph:  This could use a little editorial
wordsmithing:      For one, "SYSLOG over DTLS over DCCP [RFC5238
<http://tools.ietf.org/html/rfc5238> ]" - is it clear where the
parantheses are set?  Might consider putting the reference into the
previous paragraph ("DTLS has been mapped onto different transports").
Actually to that last sentence in the second-to-last-paragraph, is it
true that DTLS was mapped onto different transports specifically just to
secure syslog?  This is what it sounds like.  

 

3. In addition, the Introduction also states: 
"For systems where DCCP is either not available or not

   usable (such as the aforementioned situation), DTLS over UDP is also

   defined. "  

At the same time, section 5.1 states:

"Implementations of this

   specification MUST support DTLS over UDP"

 

So, the statement in the Introduction seems to be a bit misleading as it
appears to imply that DTLS over UDP is optional, specifically as it
seems to make the decision whether or not to implement it dependent just
on what is suppoted on the system (and not end-to-end considerations).  

 

4. I am not sure about the purpose of the Introduction's last sentence.
("Syslog over TLS does not
   provide application layer acknowledgements and therefore is not a
   fully reliable solution.")  If anything, this seems to belong into
the second paragraph, where it talks about performance as an issue that
is part of the motivation for a different transport.  
 
5. Section 5.1: the term "session" should be introduced.  This is the
first time the term occurs in the document.  What is the relevance of a
session in the context here?  
 
6. Section 5.1: Last paragraph, first sentence can use wordsmithing
("When TCP is used syslog over DTLS MUST NOT be used.")  When TCP is
used for what?  Might better state that syslog over DTLS must only be
used when DTLS does not use TCP.  Actually, why is this prohibition
there in the first place - is it simply not a good idea, or must it
really be prohibited?  
 
7. Section 5.1, 2nd sentence: needs wordsmithing - for one, has-->have;
is this really a single port?  
 
8. Section 5.4.1: I think there is a little potential for confusion
regarding message length.  I am assuming that message length refers to
the length of the syslog message, per section 5.4.  5.4 also states that
syslog messages do not have to align with DTLS records - allowing
application data presumably to be fragmented across frames (as a syslog
message is always contained as a whole within a syslog frame).  However,
in 5.4.1, it is also stated that "The message size SHOULD NOT exceed the
DTLS maximum record size limitation".  Why is that?  (And, is "message
size" the same as "(syslog) message length".  
 
Kind regards
--- Alex

 

_______________________________________________
Syslog mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/syslog

Reply via email to