On Wed, Mar 29, 2017 at 10:25:31AM -0700, Andy Zhou wrote: > On Fri, Mar 17, 2017 at 8:16 AM, Ben Pfaff <[email protected]> wrote: > > Signed-off-by: Ben Pfaff <[email protected]> > > --- > > Documentation/faq/openflow.rst | 8 ++++++++ > > 1 file changed, 8 insertions(+) > > > > diff --git a/Documentation/faq/openflow.rst b/Documentation/faq/openflow.rst > > index 376e64eb4482..214e38e6a9aa 100644 > > --- a/Documentation/faq/openflow.rst > > +++ b/Documentation/faq/openflow.rst > > @@ -462,6 +462,14 @@ What's going on? > > messages and will send an error response if any other value of this > > field > > is included in a "packet-out" or a "flow mod" sent by a controller. > > > > + Packet buffers have limited usefulness in any case. Table-miss > > packet-in > > + messages most commonly pass the first packet in a microflow to the > > OpenFlow > > + controller, which then sets up an OpenFlow flow that handles remaining > > + traffic in the microflow without further controller intervention. In > > such > > + a case, the packet that initiates the microflow is usually small, which > > + means that the switch sends the entire packet to the controller and the > > + buffer only saves a small number of bytes in the reverse direction. > > + > I can see that In case of TCP, the packet buffer is usually small > (except DOS). But > it is not clear to me that why this is true in general. May be you > mean this is what we > have observed in practice?
Yes. Here's a version that is slightly reworded: --8<--------------------------cut here-------------------------->8-- From: Ben Pfaff <[email protected]> Date: Wed, 29 Mar 2017 11:25:38 -0700 Subject: [PATCH] faq: Expand on answer about packet buffering removal. Signed-off-by: Ben Pfaff <[email protected]> --- Documentation/faq/openflow.rst | 9 +++++++++ 1 file changed, 9 insertions(+) diff --git a/Documentation/faq/openflow.rst b/Documentation/faq/openflow.rst index 376e64eb4482..d21bb4459395 100644 --- a/Documentation/faq/openflow.rst +++ b/Documentation/faq/openflow.rst @@ -462,6 +462,15 @@ What's going on? messages and will send an error response if any other value of this field is included in a "packet-out" or a "flow mod" sent by a controller. + Packet buffers have limited usefulness in any case. Table-miss packet-in + messages most commonly pass the first packet in a microflow to the OpenFlow + controller, which then sets up an OpenFlow flow that handles remaining + traffic in the microflow without further controller intervention. In such + a case, the packet that initiates the microflow is in practice usually + small (certainly for TCP), which means that the switch sends the entire + packet to the controller and the buffer only saves a small number of bytes + in the reverse direction. + Q: How does OVS divide flows among buckets in an OpenFlow "select" group? A: In Open vSwitch 2.3 and earlier, Open vSwitch used the destination -- 2.10.2 _______________________________________________ dev mailing list [email protected] https://mail.openvswitch.org/mailman/listinfo/ovs-dev
