RE: FW: [Diffserv] Informal WG last call for draft-ietf-diffserv-new-terms-06.txt

["Andrea Westerinen" <[e-mail]>]

Pulling everything up here.  It looks like #4 needs a bit more discussion -
however, it is not a major issue for me.

1. PolTerm supports SLAs as contractual - Agreed to document this alignment.
2. Dan said, "I'd be hesitant to alter the Diffserv definition, since this
was subject of a great deal of working group debate.  However, some words to
the effect that
you suggest would be appropriate."
	- This is all that I was suggesting.  I totally understand how painful
these things can be - and I think that the definitions are aligned.  PolTerm
is just broader and this was intentional.
3. Normative reference to PolTerm - OK.
4. Service provisioning policy "defines and manages to parameters and range
of values" in an SLS.  Dan asked how rules can manage to parameters and
ranges.
	- This really depends on your rules.  For static "preconfiguration" rules,
then all that you are doing is defining what you want, and making it so.  In
the more general case of provisioning rules, you might have
"preconfiguration" rules, and other "runtime" rules that are invoked "if
<something happens>."  If/when something happens, then the actions of the
other rules might manipulate the environment to maintain (ie, "manage to")
the parameters and ranges, or switch to new parameters, or ignore
parameters/ranges, etc.  I was asking about whether we should generalize
what a "provisioning rule" might do.
5-7. Nits and typos - Will be cleaned up.

Andrea

-----Original Message-----
From: [e-mail] [mailto:[e-mail]]On Behalf
Of Dan Grossman
Sent: Wednesday, November 14, 2001 8:09 AM
To: Andrea Westerinen
Cc: Diffserv@Ietf. Org; Policy@Ietf. Org; Bert Wijnen (Bert)
Subject: Re: FW: [Diffserv] Informal WG last call for
draft-ietf-diffserv-new-terms-06.txt


Andrea,
Thanks for the careful review.
> Dan, I was asked to do a review of the "new-terms" draft wrt PolTerm.  In
> general, there is good correspondence, but I have a few minor comments.
>
> 1. In Section 2, in the second paragraph, it would be helpful to reference
> the Policy Terminology RFC XXX (draft-ietf-policy-terminology-04.txt,
> waiting in the editor's queue) explicitly, since that supports your
> conclusions of SLAs being of a contractual nature.

Agreed.

>
> 2. The SLS definition in "new-terms" is a bit different (more restrictive)
> than the definition in PolTerm.  The PolTerm definition tries to straddle
> the contractual versus service/conditioning aspects of an SLS.  PolTerm
> states that an SLS ...
>         Specifies handling of customer's traffic by a network
>         provider. It is negotiated between a customer and the
>         provider, and (for example) in a DiffServ environment,
>         defines parameters such as specific Code Points and the
>         Per-Hop-Behavior, profile characteristics and treatment of
>         the traffic for those Code Points. An SLS is a specific SLA
>         (a negotiated agreement) and its SLOs (the individual
>         metrics and operational data to enforce) to guarantee
>         quality of service for network traffic.
> It might be valuable to mention that the "new-terms" definition further
> restricts that specified in PolTerm, for the DiffServ environment.

I'd be hesitant to alter the Diffserv definition, since this was subject of
a
great deal of working group debate.  However, some words to the effect that
you suggest would be appropriate.
>
> 3. In the last paragraph of Section 2, where you quote from the PolTerm
RFC,
> could you make this a normative reference?

Ok. I'd lost track of the Polterm draft, and frankly forgot to check the RFC
Editors queue for its status.  The way this was done reflected a concern
that
this draft would get hung up behind the Polterm draft indefinitely.  Since
the
Polterm draft is ahead, it makes perfect sense to make the Polterm RFC
normative.

>
> 4. In the last sentence of Section 2, you say "Therefore, the relationship
> between an SLS and a service provisioning policy is that the latter is, in
> part, the set of rules that define the parameters and range of values that
> may be in the former."  Would it be reasonable to say "... define and
manage
> to the parameters and range of values ..."?  The word "define" seems too
> restrictive.
>
I'm having problems parsing your proposed sentence.   How can a set of rules
"manage to" parameters and and a range of values?

> And, a few nits:
> 5. There are several weird quotation marks in the Status (``work in
> progress'' and ``1id-abstracts.txt'').
There's some wierd character mappings in my HPUX workstation.  I'll move it
back to a windows platform and spot clean when this goes to the RFC editor.

>
> 6. In Section 6, there are a few typos...
> 'Hence the imperitive was"SHOULD" rather than "SHALL"' - should be
> "imperative" and needs a space after "was"
oops...

> 'An egress DS-node at the edge of one DS-domain forwards packets an
ingress
> DS-node at the edge of another DS domain.' - needs a "to" after "forwards
> packets"

well spotted.
>
> 7. Is there a reason for the last paragraph of Section 8 to look different
> than the previous RFC summaries?  The grammar is written to correspond,
but
> the missing ":" makes the paragraph difficult to read.

Will look into, thanks.
>
> Thanks for considering these.

Thanks again for the review.
> Andrea
>
Dan


_______________________________________________
diffserv mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/diffserv
Archive:
http://www2.ietf.org/mail-archive/working-groups/diffserv/current/maillist.h
tml



_______________________________________________
diffserv mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/diffserv
Archive: http://www2.ietf.org/mail-archive/working-groups/diffserv/current/maillist.html