Re: [nmrg] Updating draft-irtf-nmrg-autonomic-network-definitions-00

["Michael Behringer (mbehring)" <[e-mail]>]

Sheng, 

> -----Original Message-----
> From: Sheng Jiang [mailto:[e-mail]]
> Sent: 27 June 2014 03:33
> To: Michael Behringer (mbehring); [e-mail]; [e-mail]
> Subject: RE: Updating draft-irtf-nmrg-autonomic-network-definitions-00
> 
> >> One comments: the current definition of autonomic network looks
> vague.
> >> May I propose to modify it: a network which employs autonomic
> >> functions network-wide.
> >
> >Thanks for the suggestion Sheng. Well, I think the definition is
> >recursive, and actually DOES reflect what you're suggesting, because it
> says:
> >
> >      <t>Autonomic Network: A network containing autonomic nodes.</t>
> >and
> >      <t>Autonomic Node: A node which employs autonomic functions. It
> >may
> >      operate on any layer of the networking stack. Examples are routers,
> >      switches, personal computers, call managers, etc.</t>
> >
> >The only thing that's not covered is the "network wide". But that's
> >also not necessarily true, because an autonomic function could well
> >span only a few devices.
> 
> Hi, Michael,
> 
> The point is how much scale of autonomic function can be called an
> autonomic network. Say, if there is a network with hundreds devices, two of
> these devices has been deployed some autonomic functions. Can the
> network be claimed as an autonomic network? "Network wide" does not
> mean every devices. It means a large portion of the network have
> autonomic functions.

The fundamental question is whether we want to include autonomic functions that for example just involve two devices. And I would strongly suggest YES, for example I can see improvements on VRRP in the future as an autonomic function (you could already claim some limited autonomy, in fact). We want to cover that, right? 

Michael

> >We could change the Autonomic Function definition to (last sentence
> added):
> >
> >      <t>Autonomic Function: A function which requires no
> >configuration, and
> >      can derive all required information either through
> >self-knowledge, discovery
> >      or through intent. Typically an autonomic function spans several
> >nodes.</t>
> >
> >What do you think? I'm a bit in between "yes, this is more precise",
> >and "this doesn't add much value and complicates things". Slight
> >tendency to agree to add it.
> 
> This is helpful. But it does not address my concern regarding to the
> autonomic network definition.
> 
> Regards,
> 
> Sheng
> 
> >Thoughts?
> >
> >Michael
> >
> >
> >>
> >> Best regards,
> >>
> >> Sheng