Hi Anil,
Some steps:
1.
Start the controller:
For eg. if you are starting NOX, in the nox server
start
./nox_core -i ptcp:6634 where 6634 is the port number
Similarly you can start floodlight, which I guess you can find in the
floodlight tutorial page.
2. Say you are doing the protocol based virtualization given in the
flowvisor tutorial(which has 2 switches),
[i.e. there are 6 hosts and all 3 pairs of hosts are connected to a switch
pair but each pair of hosts will either do a ssh or http or a telnet
connection using the same two switches ]
then you can execute the foll. commands:
Start flowvisor:
./flowvisor.sh &
*# Creation of 3 Slices*
fvctl createSlice SSH tcp:127.0.0.1:6633 <email id> ----> Here instead
of 127.0.0.1, you can specify the controller's ip address(127.0.0.1 is the
local host if you are running it locally)
fvctl createSlice HTTP tcp:127.0.0.1:6634 <email id>
fvctl createSlice TELNET tcp:127.0.0.1:6635 <email id>
*# We must add ARP traffic to all slices*
fvctl addFlowSpace any 100 dl_type=0x806,dl_src=00:00:00:00:00:0A,nw_dst=
192.168.1.0/24 Slice:SSH=4 ---> dl_src matches Ethernet source address mac
which should correspond to the host mac address
connected to one of the switches in your case;
dl_type is ethernet protocol type and 0x0806 matches ARP packets;
nw_dst is the destination host ip address to which you are trying to do
a 'SSH' in this slice;
SSH=4 is ssh slice with write permission.
fvctl addFlowSpace any 100 dl_type=0x806,dl_src=00:00:00:00:00:0D,nw_dst=
192.168.2.0/24 <http://192.168.1.0/24> Slice:SSH=4
fvctl addFlowSpace any 100 dl_type=0x806,dl_src=00:00:00:00:00:0B,nw_dst=
192.168.3.0/24 <http://192.168.2.0/24> Slice:HTTP=4
fvctl addFlowSpace any 100 dl_type=0x806,dl_src=00:00:00:00:00:0E,nw_dst=
192.168.4.0/24 <http://192.168.2.0/24> Slice:HTTP=4
fvctl addFlowSpace any 100 dl_type=0x806,dl_src=00:00:00:00:00:0C,nw_dst=
192.168.5.0/24 <http://192.168.3.0/24> Slice:TELNET=4
fvctl addFlowSpace any 100 dl_type=0x806,dl_src=00:00:00:00:00:0F,nw_dst=
192.168.6.0/24 <http://192.168.3.0/24> Slice:TELNET=4
*# Now we can add FlowSpace for the actual traffic.*
fvctl addFlowSpace any 100 dl_type=0x800,nw_proto=6,nw_src=
192.168.1.0/24,tp_dst=22 Slice:SSH=4 ---> Here nw_proto=6 means TCP packets
fvctl addFlowSpace any 100 dl_type=0x800,nw_proto=6,nw_src=
192.168.1.0/24,tp_src=22 Slice:SSH=4
fvctl addFlowSpace any 100 dl_type=0x800,nw_proto=6,nw_src=
192.168.2.0/24,tp_dst=22 <http://192.168.1.0/24,tp_dst=22> Slice:SSH=4 --->
Here nw_proto=6 means TCP packets
fvctl addFlowSpace any 100 dl_type=0x800,nw_proto=6,nw_src=
192.168.2.0/24,tp_src=22 <http://192.168.1.0/24,tp_src=22> Slice:SSH=4
----> Here you are adding flowspace for all the traffic being generated
from and received at the port 22(ssh port) of 192.168.1.0/24 and
192.168.2.0/24 hosts
respectively so that most ssh traffic scenarios between these two hosts get
covered.
Similarly you have to add flowspaces for the other two slices viz.
HTTP(port 80) and TELNET(port 23) with the respective host ip addresses.
Try them out. Please do some tweaking of the above steps if they do not
work as it is.
Else you can do slicing for any other topology in any other
manner(topology/protocol based) you want,not necessarily the above way but
the
fundamental steps are as above itself(i.e. create slices, add arp flowspace
with dl_type=0x806 and add the actual traffic with their src and dst ip
address)
Best Regards
Sumanth
On Sat, Mar 9, 2013 at 3:09 AM, anil4haart <anil4ha...@gmail.com> wrote:
> Hi ,
>
> I want three scenarios - which involves :
>
> 1. Controller ( FloodLight preferably) .
> 2. Two switches ( Physical Switches ).
> 3. FlowVisor ( Running on the different machine ).
>
> *The main Idea for this is :*
> I want to use my testing framework in the openflow enabled devices:
>
> The main steps will be :
> - Configure the Devices and start the controller etc.
> - Add the flowSpace to the slice.
> - Verify the configuration and slicing.
>
> I am not that much aware of the internal functionality of the devices , It
> would be great if you give clear description.
>
> Thanks,
> Anil
>
> _______________________________________________
> openflow-discuss mailing list
> openflow-discuss@lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/openflow-discuss
>
>
_______________________________________________
openflow-discuss mailing list
openflow-discuss@lists.stanford.edu
https://mailman.stanford.edu/mailman/listinfo/openflow-discuss
