Re: [Netconf] review of draft-ietf-netconf-with-defaults-03
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Netconf] review of draft-ietf-netconf-with-defaults-03



Hello,
A late comment.

The cost of storage needed to store whether a leaf was set explicitly is usually cheap, BUT the cost of adapting an existing configuration datastore and configuration handling logic is EXPENSIVE.

So if we would make explicit mandatory for Netconf would significantly raise adaptation costs for many existing devices.

Balazs

On 09/25/09 17:57, Phil Shafer wrote:
Andy Bierman writes:
It is wrong to assume
that NV-storage of 1 XML file is somehow a burden for
every platform running NETCONF.

Well, it's wrong to assume that a platform running NETCONF uses XML
for anything more than parsing content off the wire.  The JUNOS
config database is not XML, nor are our NV-storage contents.

Also the storage needs for "explicit" or "trim" are not similar to
"report-all".  For example, the JUNOS list element for static routes
has nearly 75 leafs under it, of which 2-3 are typically set.
Assuming they all have default (which they don't, but...), the
customer that loads 50,000 static routes has a impact (with 32 bytes
overhead per list instance plus, say, 12 bytes per leaf instance)
of 2.8M for explicit and 46.6M for report-all.

Thanks,
  Phil
_______________________________________________
Netconf mailing list
Netconf at ietf.org
https://www.ietf.org/mailman/listinfo/netconf


--
Balazs Lengyel                       Ericsson Hungary Ltd.
System Manager
ECN: 831 7320                        Fax: +36 1 4377792
Tel: +36-1-437-7320                  email: Balazs.Lengyel at ericsson.com

Note: Messages sent to this list are the opinions of the senders and do not imply endorsement by the IETF.