[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: RDF ( minutes for NETCONF WG interim meeting (09/03))
> [In fact, I liked the RDF inspired exercise to view a whole device as
> a collection of resources where each resource is identified by a URI
> so the whole instance naming would be based on URIs. I am not
> suggesting to do this in netconf - but it is an interesting approach
> to think through. Perhaps this is what you would call "anarchy" since
> we would no longer dictate how a vendor has to identify its
> interfaces? But perhaps this is not even bad since we know from
> experience that getting indexing right is hard and we went to
> introduce and use contexts in SNMP land for those cases where we
> screwed up. And contexts per se are rather opaque - I would call
> that anarchy - not sure that matches your definition.]
>
I am greatly interested in the approach of RDF to describe the network
resource, too. Don't know what the impact on the protocol is though. Maybe
nil.
--
Weijing Chen
> -----Original Message-----
> From: Juergen Schoenwaelder [mailto:schoenw@ibr.cs.tu-bs.de]
> Sent: Thursday, October 16, 2003 12:04 PM
> To: randy_presuhn@mindspring.com
> Cc: netconf@ops.ietf.org
> Subject: Re: minutes for NETCONF WG interim meeting (09/03)
>
>
> >>>>> Randy Presuhn writes:
>
> >> I think the key point is the need for parsing the identifier. We
> >> don't need to parse OIDs. They are good as global identifiers via
> >> decentralized naming authorities.
>
> Randy> That's what they're supposed to be, and how the SMI uses them,
> Randy> but that's not how SNMP uses them. Recall what it takes to
> Randy> extract table indexes from object identifiers. Subagents spend
> Randy> much of their time parsing OIDs, and we've heard complaints
> Randy> from application writers who wish index objects were accessible
> Randy> so they wouldn't have to parse OIDs as frequently.
>
> This thread is about naming capabilities and so the strange way SNMP
> uses to name instances does not really apply to this discussion.
>
> Randy> A bystander's comment: This argument about structure
> Randy> vs. anarchy for the substance of this protocol seems to recur
> Randy> for every element, and the bias (perhaps due to the charter)
> Randy> seems to be towards anarchy. I think this will undercut the
> Randy> value of whatever eventual standardization of those elements
> Randy> might happen.
>
> Not sure I understand this remark. Are you saying that using URIs
> for identifying capabilities is "structure" or "anarchy"?
>
> [In fact, I liked the RDF inspired exercise to view a whole device as
> a collection of resources where each resource is identified by a URI
> so the whole instance naming would be based on URIs. I am not
> suggesting to do this in netconf - but it is an interesting approach
> to think through. Perhaps this is what you would call "anarchy" since
> we would no longer dictate how a vendor has to identify its
> interfaces? But perhaps this is not even bad since we know from
> experience that getting indexing right is hard and we went to
> introduce and use contexts in SNMP land for those cases where we
> screwed up. And contexts per se are rather opaque - I would call
> that anarchy - not sure that matches your definition.]
>
> /js
>
> --
> Juergen Schoenwaelder International University Bremen
> <http://www.eecs.iu-bremen.de/> P.O. Box 750 561, 28725 Bremen,
> Germany
>
>
>
> --
> to unsubscribe send a message to netconf-request@ops.ietf.org with
> the word 'unsubscribe' in a single line as the message text body.
> archive: <http://ops.ietf.org/lists/netconf/>
--
to unsubscribe send a message to netconf-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/netconf/>