[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <4ED3ADCC.604@nicira.com>
Date: Mon, 28 Nov 2011 07:50:36 -0800
From: Martin Casado <casado@...ira.com>
To: Jamal Hadi Salim <jhs@...atatu.com>
CC: Herbert Xu <herbert@...dor.apana.org.au>, dev@...nvswitch.org,
netdev@...r.kernel.org, David Miller <davem@...emloft.net>
Subject: Re: [ovs-dev] [GIT PULL v2] Open vSwitch
>> This is a common misunderstanding about OpenFlow. It does not require
>> the first packet of each flow to go to the controller.
> I am reading this to mean that the switch CPU will resolve things?
> Typically those tend to be tiny cpus.
>
No, no datapath traffic leaves the switching chip. Generally
controllers calculate full forwarding tables using wildcarded rules and
use low priority rules to handle exception traffic (still keeping it on
the datapath).
>
>> In fact, no
>> production system I'm aware of does this. Generally OpenFlow-based
>> solutions targeted at large environments (e.g. data center, or WAN)
>> send only traditional control traffic to the controller (e.g. BGP or
>> OSPF), or none at all.
> Even OSPF or BGP would be problematic imo if the architecture doesnt
> allow prioritization of some form towards the controller.
Indeed. Controllers generally implement such prioritization by using
OpenFlow forwarding rules that match on control traffic and explicitly
set the controller as the destination. This allows the application of
any QoS primitives within OpenFlow to control traffic.
best,
.martin
>
> cheers,
> jamal
>
>
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~
Martin Casado
Nicira Networks, Inc.
www.nicira.com
cell: 650-776-1457
~~~~~~~~~~~~~~~~~~~~~~~~~~~
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists