Hi Jonathan,
These are just my thoughts ...
I'm not advocating getting rid of candidate/locking, but if I was
designing a new network management solution, I would not plan on using
candidate/locking.
Candidate is great when you have got a human operator typing in some
config on the CLI and want to stop anyone (or anything) else changing
the config at the same time under their feet.
But once you only have machines programming the config, then they should
be able to easily construct a single message for each edit of the config
(which can be applied/failed in its entirety). Candidate and locking
don't seem to help here.
Further, if you want to update multiple devices at the same time then
you end up in the realm of distributed transactions which get very
complicated and are hard to get right in a fully robust fashion.
It is at this point, that I would try hard to build a network management
architecture that doesn't rely on transactions or locking, and instead
only relies on individual updates being consistently applied in a
serialized fashion.
Rob
On 19/12/2016 08:09, Jonathan Hansford wrote:
-----Original Message-----
From: netmod [mailto:[email protected]] On Behalf Of Ladislav
Lhotka
Sent: 14 December 2016 14:25
To: Andy Bierman <[email protected]>; Mehmet Ersue
<[email protected]>
Cc: NetMod WG Chairs <[email protected]>; NetConf WG Chairs
<[email protected]>; NetMod WG <[email protected]>; Netconf
<[email protected]>
Subject: Re: [netmod] [Netconf] WG adoption poll draft-nmdsdt-netmod-
revised-datastores-00
:
As for candidate, it is optional and we all know that it is quite
problematic if
concurrent access of multiple clients is possible. Therefore, it would IMO
be a
good riddance.
For someone who is yet to see NETCONF and YANG used in anger, can you
explain why judicious use of candidate and lock is problematic with
concurrent access and why, as a consequence, it should be got rid of? One of
the features of NETCONF that attracted us to it is its support for
transactioning, a feature which it would appear came from previous
experience with JUNOS. Is it current guidance that transactioning should not
be used?
Jonathan
Lada
_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod
.
_______________________________________________
netmod mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/netmod
