>>>>> On Wed, 3 Feb 2010 16:37:38 -0500, "David Harrington" 
>>>>> <[email protected]> said:

[cut to just the ISMS document's wording:]
DH> DTLS Transport Model server implementations MUST support DTLS
DH> cookies.

DH> Implementations are not required to perform the stateless cookie
DH> exchange for every DTLS handshake, but in environments where an
DH> overload on server side resources is detectable by the
DH> implementation it is RECOMMENDED that the cookie exchange is
DH> utilized by the implementation.

[And your comments:]
DH> My impression is that syslog allows an admin to enable this as a
DH> deployment option; tlstm seems to depend on the implementation. (and I
DH> have difficulty knowing how an implementation will know if overlaod on
DH> server side is detectable, unless it is doing the detection itself.)
DH> 

Actually, we require it to be implemented but we're rather flexible on
how it's actually used.  Intentionally.  I decided that there wouldn't
be problems if different implementations did different things.  Some
implementations could have it configurable; others could do automatic
detection, etc.  So we certainly support the same case that syslog is
doing, but we also allow implementations more flexibility.  I don't
believe this impacts interoperability of our base protocol.

DH> 1) Is enable/disable capability a MUST implement for DTLS cookie
DH> exchange?

Not currently.

DH> 2) Should tlstm recognize whether it is enabled/disabled? 

That's diving into protocol layers we said we wouldn't dive into.
Specifically, I was requested not mention DTLS specific protocol message
types in the document (like the syslog/DTLS document does).

DH> 3) Would it be better to RECOMMEND or REQUIRE the cookie exchange in
DH> syslog when the implementation can detect an overload on server
DH> resources?

Then that would prohibit an implementation from doing a runtime
on/off/maybe setting.  So personally, no, I don't see the need.  But I
also don't feel strongly about it.

-- 
Wes Hardaker
Cobham Analytic Solutions
_______________________________________________
Syslog mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/syslog

Reply via email to