Re: Scalability as requirement

[Andy Bierman <ietf@andybierman.com>]

Juergen Schoenwaelder wrote:
> On Thu, May 18, 2006 at 01:45:18PM +0200, Balazs Lengyel wrote:
> > Jurgen,
> > Could you please add scalability to your requirements page. After the 
> > discussions I believe we require that the protocol should allow 30.000 to 
> > 100.000 nodes to send notifications to a management station and the station 
> > should be able to order the 100.000 nodes to send the notifications.

This is a really poorly worded requirement.
A million nodes could send notifications.
The protocol does not actually constrain the number
of nodes that can participate in netconf sessions.
Nor would it.

Operational scalability is hard to quantify, and it
of course depends on your acceptable failure response time
and security requirements.  I don't want to specify maximum
event-to-delivery latency times in the netconf standard.
I don't know how to write this requirement in a meaningful
and acceptable way.


IMO, Callhome could sufficiently address the scalability problem
wrt/ manager-subscribe based delays, if it was fully specified.
The relative overhead of establishing a netconf session to send
secure notifications is debatable I guess, about the same any
way you do it (if you decide up front you want to use sessions).


> Added to the list.
>
> /js

Andy

--
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/>