Hi Murphy:
    Thank you very much for your detailed reply.  The PC1 and PC2 can ping each 
other successfully when they are on the same subnet. The layer 3 proceesing 
which i refered last email means that the PC1 and PC2 can also ping each other 
successfully when they are not on the same subnet. In fact, they can not 
communicate with each other on my test network when on different subnet. 
    At last,i wonder whether attachments can be received by your mailbox. If 
they can, i'll send them with the email together.
Best Wishes 
                                                                                
                                                      Wang Jian
                                                                                
                                         Tuesday, December 6,2011



在 2011-12-06 11:04:04,"Murphy McCauley" <jam...@nau.edu> 写道:
Responses inline.


On Dec 5, 2011, at 6:42 PM, 王健 wrote:


Hi Murphy:
    I am confused about a question about OpenFlow dedicated switch for a long 
time.The following is my testing network topology:
                                      <截图4.png>
PC1 and PC2 can ping each other successfully if they have the same net id, and 
can't ping successfully if they don't.


I assume you mean if they're on the same subnet?


Does OpenFlow dedicated switch only surpport layer 2 processing when run the 
command : nox_core -v -i ptcp:6633 discovery topology authenticator routing?


Well, routing tries to install exact match flows whenever possible, if I recall 
correctly.  This includes L3 headers too when possible.  But it's not exactly 
L3 *processing*.


I think it has no ability of forwarding if disconnects to NOX.


There are things that can be done to help with communication when NOX is 
disconnected, but none of them are done here.  Certainly the usual model is 
that your controller stays connected to your switches.


I just want to know how to achieve layer 3 processing with OpenFlow dedicated 
switch and NOX. Would you please help me even some advice,thank you very much~


What kind of layer 3 processing are you talking about?


   Another question,how to make the topo update timely which you can see from 
NOX GUI , the topo can display the whole network topo when run the command: 
./nox_core monitoring discovery topology trackhost_pktin. But it can't update 
timely when a host leave from the network. How to solve this problem? I am 
sorry to post it again as the question may be overloaded. Thank you~


Well, there's no real way to tell when a host leaves.  So right now, it's 
decided by timeout.  You could try reducing the timeout, but that increases the 
chance that a quiet host may be judged to be leaving when it's actually just 
quiet.


I have suggested that we ARP-ping hosts occasionally to see if they're still 
there or not.  If a host NEVER responds to an ARP-ping, fall back to the 
timeout behavior.  This hasn't yet been implemented, though, as far as I know.


    By the way, i have solved the disconnecting problem via adding a 1G 
ethernet switch beween them yesterday.Because the netfpga only surpports 1G 
transporting speed,but the max speed which the openwrt surpports is 100M.
Best Wishes
                                                                                
                                          Wang Jian
                                                                                
                                 Tuesday, December 6,2011


-- Murphy

_______________________________________________
nox-dev mailing list
nox-dev@noxrepo.org
http://noxrepo.org/mailman/listinfo/nox-dev

Reply via email to