[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RDF and network configuration



At 23:29 14/03/04 +0100, Juergen Schoenwaelder wrote:
On Thu, Mar 11, 2004 at 09:51:04AM +0000, Graham Klyne wrote:

> Hmmm... I find the description of RDF schemas as expressing *grammars* is a
> little odd, but I suppose it's close enough. I do think that RDF would be
> a good choice for the data model used to express device characteristics,
> well suited to dealing with evolving requirements for device description,
> and supported by tools that can help one reason about system configurations
> (see also the work on OWL). (I've done a little work to generate device
> configuration files from RDF.)


I think the key is the naming scheme. By adopting URIs as the naming
system for all the object instances on a device, RDF can naturally
be used. (But I think we would still need quite some XML schema
definitions to describe the structure of complex objects.)

I agree about the key role of URIs.


RDF descriptions can have as complex a structure as you need. One way to use RDF schema, especially in conjunction with tools like OWL, is to verify that required structural conditions are satisfied. (In this respect, RDF schema might be regarded as describing a "grammar", but there is a "missing isn't broken" [1] philosophy to RDF that allows it to be more flexible and evolvable than a fixed XML schema. It seems to me that this open-ended-ness is ideally suited to describing network devices, whose future characteristics can only be guessed at.

Figuring out how best to combine RDF descriptions with XML document schema is an area of considerable ongoing research interest. There seems to be an emerging consensus that XSLT will have a key role to play [3].

Anyway, since you seem to have concrete experience, I think it would be
nice if you can post a more concrete description of what you are doing
so we all can read and learn.

Well, it was a small amount of work I did about a year ago, a demonstration to enforce access restrictions in my home network. Some links to my write-up and notes can be found at [2]. Since then, the Cisco router I was using has died, but I do still use a dhcp.conf and a DNS zone file that are both generated from a common RDF description.


My work on this was discontinued when some budgets were re-allocated, and I've been working separately on other aspects of Semantic Web technology, but I do still have a goal of revisiting this topic to generate a Linux firewall (IPChains) configuration from RDF policy descriptions. I think that using RDF to describe network devices would make this kind of project far easier, and of considerable value.

#g
--

[1] http://rdfweb.org/mt/foaflog/archives/000047.html
Dan Brickley,
"Missing isn't broken: data validation and freedom on the Semantic Web"

[2] http://www.ninebynine.org/SWAD-E/Intro.html#HomeNetAccessDemo
has a series of links describing aspects of my small experiments performed as part of the European Semantic Web Advanced Development (SWAD-E) project to generate access control configuration files from a policy description. A keyn thrust of this experiment was to use *only* existing RDF tools, not writing new software.


[3] http://www.w3.org/2001/sw/interest/meetings/tp2004.html
Recent meeting of W3C Semantic Web Interest Group, included discussion of:
http://www.w3.org/MarkUp/2004/02/xhtml-rdf.html
http://www.w3.org/2004/01/rdxh/spec
http://www.hpl.hp.com/techreports/2003/HPL-2003-268.html



------------
Graham Klyne
For email:
http://www.ninebynine.org/#Contact


-- 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/>