[openflow-discuss] interaction between openflow and routing (and possibly upper layers)

2012-05-23 Thread Daniel Luo
Hello folks, Please correct my understanding below if I'm wrong: For the 1st packet of a flow, the controller will need to set up the route for that packet. To do this it will resort to an external routing protocol which is outside the scope of openflow. Let's say the protocol is a link state rou

Re: [openflow-discuss] OpenRoads controller link with PC Engine AP

2012-05-23 Thread Heming Wen
Hello again, First of all, I tried L2 inband mode (APTYPE=2) with VLAN and the AP does not disconnect from the controller. However, with VLAN I can't access the rest of my network so I would still like to use L3 inband (APTYPE=0). Thus, I looked further in the L3 inband configuration and notice

Re: [openflow-discuss] Openflow ncasting Client

2012-05-23 Thread kk yap
Hi Kevin, I am suggesting using OVS to bond the interfaces. And Ben is right too. :) How about changing the source mac address of the wireless interfaces to be the same using ifconfig? This will get around the problem. Of course, not all interfaces allow you to do that, but plenty does. If

Re: [openflow-discuss] Openflow ncasting Client

2012-05-23 Thread Kevin Han
Hi KK, I have tried to use Open v switch to replace Hoolock by attempting to bond two wireless NICs together. However, after many unsuccessful attempts, I was informed that bridging between wireless NICs does not work in general (from Ben). I thought your original suggestion about using OVS was to

Re: [openflow-discuss] Wireless NIC Bonding for n-casting demo

2012-05-23 Thread Ben Pfaff
On Wed, May 23, 2012 at 04:15:20PM -0400, Kevin Han wrote: > 1) When I use openvswitch to create a bond (i.e. ovs-vsctl add-bond br0 > bond0 eth0 eth1 lacp=active trunks=10,20) the lacp-negotiated value is > false and the bond does not work. Someone from ovs mailing told me that it > doesn't work

[openflow-discuss] Wireless NIC Bonding for n-casting demo

2012-05-23 Thread Kevin Han
Hello, I am trying to re-create the openflow wireless n-casting demo but I am having a bit of trouble with the client. I have two wireless NICs (one internal Intel and an atheros external card) that I am using in a virtual machine via a bridge adapter (so they appear as eth0 and eth1 respectively)

Re: [openflow-discuss] oftest on remote software switch

2012-05-23 Thread Dan Talayco
Assuming you're using veth interfaces to connect the parts together, you should use local.py (which doesn't really do anything in the default case). See the references to veth in oftest/tests/oft; you may need to use --base-if-index. -Dan On Wed, May 23, 2012 at 1:11 AM, sonny sonny wrote: >

[openflow-discuss] oftest on remote software switch

2012-05-23 Thread sonny sonny
Hi, I am running OVS(software switch) on my host 10.0.15.1 (Linux ) and oftest is run on virtual machine 10.0.15.2 (Ubuntu 10.0.4) on same host. So should I run oftest using local.py or remote.py ? If I use remote.py then how to map the ports between software switch and VM(oftest). regards son