Randall Gellens wrote:
Thanks for reviewing it.
Randy: No problem; my pleasure. As always, I learn more
from reviewing ;-)
More inline.
The intent of S4 is to mandate support for TCP. The intent of the
webmail/HTTP text is to explain that, if TCP can't be supported, while
some form of mail can be offered, it isn't lemonade. Perhaps this can
be made more clear by changing the start of that paragraph from "For
environments" to "Note that, for environments", and changing the end
from "must also be limited." to "must also be limited, and can't be
lemonade compliant."
Yes, that should do it. The intent will be much more clearer
with the edits you suggest above.
2) In S4.2, would it help to have a MUST strength for
turning off TCP keepalives? More so since it is also harmful
for dormancy (as discussed in S5.)
I'm fine with that, but I'd have to ask the AD and chairs if such a
change requires an explicit blessing by the WG (if so, this could likely
be done with a quick call for consensus rather than a new WGLC and IETF
LC).
I will leave it up to you. My experience has been that implementors
pay more attention to the MUSTs than they do to the SHOULDs. And
since the text referenced by this "SHOULD NOT" potentially injects
more unwanted traffic in the Internet in the form of opening new
connections, I assumed that a MUST-level strength inhibiting this
behavior is better for the Internet at large.
3) In S6, please consider providing a reference for supporting
the sentence "There is a body of best practices based on long
experience in this area." It may help the reader if he/she
wants to explore this area in some more detail.
A reference would be nice, and I had tried to find a suitable one, but
wasn't able to find one that I felt was OK for an RFC (e.g., not a
wikipedia entry). If you have a suggestion I'd appreciate it.
A couple of places come to mind:
1) comp.security.firewalls FAQ - Take a look at S5 in there; it
talks about "How Do I..." to get traffic through a firewall.
The FAQ is archived at http://www.faqs.org/faqs/firewalls-faq/.
Having said that, I will note that the FAQ is fairly dated
(Dec 2000.) But, its maintainers are Marcus Ranum and Matt
Curtin. Of these, I have heard of Marcus' name in the security
field quite a bit, and I dare say that someone in Gen-ART or
the IETF security directorate may know him and can maybe ask
him on current literature on firewall avoidance. FWIW, Marcus'
email is provided on the FAQ page.
2) P2P literature. By its nature, P2P networks have to work
across firewalls. Liang et al. [1] discuss some techniques
(very briefly) in P2P networks.
[1] J. Liang, R. Kumar, K.W. Ross, "The KaZaA Overlay: A
Measurment Study," Computer Networks Journal, Elsevier, 2005.
Archived at http://photon.poly.edu/~rkumar04/KazaaOverlay.pdf.
Nits:
1) Just a suggestion: please consider expanding acronyms on
first use.
I tried to follow a practice of expanding acronyms on first use except
those that are in such wide use within the email area that doing so
might be distracting (such as TLS, POP, IMAP, SMTP), but if it makes the
document easier for non-experts to read, I'd be happy to expand others.
Are there particular ones that you suggest expanding?
No, not really. To anyone conversant in this domain, the acronyms
should be easily understood. I was more worried about non-experts.
But I am also unsure how much to bend backwards to accomodate them.
I agree with you that the ones you mention above should be well known
by anyone in the field. I also agree that expanding well-known
acronyms simply distracts. In the end, I will simply leave it to
your sound editorial discretion.
Thanks,
- vijay
--
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
2701 Lucent Lane, Rm. 9F-546, Lisle, Illinois 60532 (USA)
Email: [EMAIL PROTECTED],bell-labs.com,acm.org}
WWW: http://www.alcatel-lucent.com/bell-labs
_______________________________________________
Gen-art mailing list
[email protected]
https://www1.ietf.org/mailman/listinfo/gen-art