[support] can NEPL/UMIP support the VMN?

Romain KUNTZ kuntz at lsiit.u-strasbg.fr
Sat Aug 23 04:25:22 JST 2008


Hello,

The logs seems OK, I actually suspect a routing problem. When your MR  
is in the foreign link, here are the operations that should be done by  
the HA:

- HA echo request destined to VMN's HoA
- HA lookups in the routing table, route goes towards the tunnel  
associated to the VMN (you can get the routing table with "ip -6 route  
show table 252")
- The destination of the outer header of the tunnelled packet belongs  
to the MR's MNP, so I wonder if the HA does not have a problem here  
when your MR is in the foreign network. I guess your HA also have a  
route for the MR's MNP towards the home link interface right? Maybe  
the HA is not selecting the correct route (which should be the route  
toward the tunnel associated to the MR). You can check other routes by  
doing "ip -6 route show table main", and see what's the priority among  
routing tables with "ip -6 rule". You could check also if the HA does  
not try to send the icmp echo request via the home link.
- If the packet goes through the MR's tunnel (you could check that  
with tcpdump on the MR's tunnel interface), it should reach the VMN.  
Maybe the problem comes from the returning route where the packet  
could also be lost somewhere on the path.

As you can see, there could be many reasons why your ping does not  
succeed. Try to check where the problem is located (MR, HA, VMM), and  
if the routes are correctly selected by the HA (you could for example  
tcpdump in various places of your testbed: home network, tunneling  
interfaces of your HA, MR, VMN).

Regards,
romain


On 2008/08/20, at 9:44, Junkang Ma wrote:

> Dear all,
>
> I have repeated some of my tests. I uses three laptops. HA has three  
> interfaces: one ethnet as Home Link, two WiFi interfaces as two  
> foreign links. MR has three interfaces: one ethnet links to HA, one  
> ethnet as NEMO-link, one WiFi links to foreign link 1. VMN has two  
> interfaces: one ethnet links to MR, one WiFi links to foreign link  
> 2. The topology is shown as follows:
>
>                 
> +xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>                 
> |                                                      x
>                |(Foreign link 2: 2001:a:d: 
> 1::/64)                     x
>                |wlan1 (2001:a:d: 
> 1::1000/64)                           x
>  +-------------+------------ 
> +                                         x
>  |                          |wlan0 (2001:a:c: 
> 1::1000/64)              x
>  |            HA            |(Foreign link 1: 2001:a:c: 
> 1::/64)        x
>  |                          +---- 
> +xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx    x
>  | HoA: 2001:a:b::1000/64   |                                     
> x    x
>  |                          |                                     
> x    x
>  +-------------+------------+                                     
> x    x
>               ||eth4 (2001:a:b::1000/64)                          
> x    x
>               ||(Home link: 2001:a:b::/64)                        
> x    x
>               ||                                                  
> x    x
>               ||                                                  
> x    x
>               ||                                                  
> x    x
>               ||eth4 (2001:a:b::1/64)                             
> x    x
>  +-------------+------------+                                     
> x    x
>  |                          |                                     
> x    x
>  |            MR            |wlan0 (link to Foreign link 1)       
> x    x
>  |                          +---- 
> +xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx    x
>  | HoA: 2001:a:b::1/64       
> |                                         x
>  |                           
> |                                         x
>  +-------------+------------ 
> +                                         x
>               ||eth6 (2001:a:b: 
> 1::1/64)                               x
>               ||(NEMO-Link: 2001:a:b: 
> 1::/64)                          x
>                
> ||                                                      x
>                
> ||                                                      x
>                
> ||                                                      x
>               || 
> eth4                                                  x
>  +-------------+------------ 
> +                                         x
>  |                           
> |                                         x
>  |            VMN           |wlan1 (link to Foreign link  
> 2)           X
>  |                          +---- 
> +xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
>  | HoA: 2001:a:b::2/64      |
>  |                          |
>  +-------------+------------+
>
> The config files of HA, MR and VMN are in the attachments: HA  
> conf.txt, MR conf.txt, VMN conf.txt.
>
> The steps of the test:
>
> [step1]
> Start HA. The log is in the attachments: log of HA when starting.txt.
>
> [step2]
> MR's ethernet (eth4) links to Home Link (HA's eth4), and MR's eth4  
> is its first priority interface. Start MR. The log is in the  
> attachments: log of MR when starting.txt.
>
> [step3]
> VMN's ethernet (eth4) links to NMEO-Link (MR's eth6), and VMN's eth4  
> is its first priority interface. Start VMN. VMN's HoA is binding to  
> its NEMO-link CoA. The log is in the attachments: log of VMN when  
> starting.txt.
>
> [step4]
> HA uses ping6 to VMN's HoA (2001:a:b::2). At the same time, HA uses  
> ping6 to VMN's CoA (the NEMO-link address gotten from VMN's eth4).  
> They both work fine! No any error displayed in the log.
>
> [step5]
> MR disconnects HA's Home Link (MR's eth4 disconnects HA's eth4). VMN  
> is kept the same (still links to MR). Binding Update (BU) of MR is  
> received by HA. MR's HoA is binding to its wlan0 CoA. VMN's HoA is  
> still binding to its NEMO-link CoA.
>
> !!! PROBLEM 1 !!!
> At this time, HA can not ping6 to VMN's HoA (2001:a:b::2), but HA  
> can ping6 to VMN's CoA (the NEMO-link address gotten from VMN's eth4).
>
> The log is in the attachments: log of HA when MR handover.txt, log  
> of MR when MR handover.txt.
>
> [step6]
> Continue with step5. MR connects to HA's Home Link again (MR's eth4  
> connects HA's eth4 again). No binding of MR exists. VMN is kept the  
> same (still links to MR). VMN's HoA is still binding to its NEMO- 
> link CoA.
>
> !!! PROBLEM 2 !!!
> At this time, HA can ping6 to VMN's HoA (2001:a:b::2) and also to  
> VMN's CoA (the NEMO-link address gotten from VMN's eth4). However,  
> ICMPv6 Request and Reply go through different paths: ICMPv6 Request  
> goes from HA's eth4 to MR's eth4, but Reply goes from MR's wlan0 to  
> HA's wlan0.
>
> No any error displayed in the log.
>
>
>
> Would you mind helping me to solve these problems?
>
> Thank you very much,
>
> Junkang
>
>
>
> Quoting Romain KUNTZ <kuntz at lsiit.u-strasbg.fr>:
>
>> On 2008/08/05, at 14:49, Junkang Ma wrote:
>>> I will go to repeat the test, and send you the test log file in  
>>> the  future email.
>>
>> Yes please. If you can also try a traceroute6 from the CN to the VMN
>> (before and after the MR's handover), that would be helpful too.
>>
>> Regards,
>> romain
>>
>>> Quoting Romain KUNTZ <kuntz at lsiit.u-strasbg.fr>:
>>>
>>>> Hello,
>>>>
>>>> On 2008/08/04, at 22:55, Junkang Ma wrote:
>>>>> I am a student who is working on the NEMO-related project. I  
>>>>> have   implemented a NEMO testbed using NEPL/UMIP. It works fine  
>>>>> when  the  MNN is a LFN. However, after I add the UMIP into the  
>>>>> MNN to  make it  becomes a VMN, the communication from the HA or  
>>>>> the CN  outside the  moving network to the VMN can not be kept  
>>>>> continuous  when the  moving network moves (MR changes its  
>>>>> network interfaces  so as to  change its CoA).
>>>>
>>>> Do you mean that the VMN becomes completely unreachable after the
>>>> handover of the MR? or that the VMN disconnection is too long  
>>>> compared
>>>> to the MR's one?
>>>>
>>>>> Could you please tell me whether NEPL/UMIP can support the VMN   
>>>>> now?  If you need more details of my test, I will send anything  
>>>>> I  have,  like HA log.
>>>>
>>>> I remember I already tested this case successfully, but that was  
>>>> a long
>>>> time ago.
>>>> I need more information about your test environment: topology,  
>>>> config
>>>> files of the MR, HA and VMN, logs of the MR, HA and VMN, etc.
>>>>
>>>> Cheers,
>>>> -- 
>>>> Romain KUNTZ
>>>> kuntz at lsiit.u-strasbg.fr
>>>> LSIIT - Networks and Protocols Team
>>>> http://clarinet.u-strasbg.fr/~kuntz/
>>>>
>>>> _______________________________________________
>>>> Support mailing list
>>>> Support at ml.nautilus6.org
>>>> http://ml.nautilus6.org/mailman/listinfo/support
>>>
>>>
>>>
>>> -- 
>>> The University of Edinburgh is a charitable body, registered in
>>> Scotland, with registration number SC005336.
>>>
>>>
>>> _______________________________________________
>>> Support mailing list
>>> Support at ml.nautilus6.org
>>> http://ml.nautilus6.org/mailman/listinfo/support
>>>
>>
>> -- 
>> Romain KUNTZ
>> kuntz at lsiit.u-strasbg.fr
>> LSIIT - Networks and Protocols Team
>> http://clarinet.u-strasbg.fr/~kuntz/
>>
>> _______________________________________________
>> Support mailing list
>> Support at ml.nautilus6.org
>> http://ml.nautilus6.org/mailman/listinfo/support
>
>
>
> -- 
> The University of Edinburgh is a charitable body, registered in
> Scotland, with registration number SC005336.
>
>
> <HA conf.txt><MR conf.txt><VMN conf.txt><log of HA when  
> starting.txt><log of MR when starting.txt><log of VMN when  
> starting.txt><log of HA when MR handover.txt><log of MR when MR  
> handover.txt>_______________________________________________
> Support mailing list
> Support at ml.nautilus6.org
> http://ml.nautilus6.org/mailman/listinfo/support

-- 
Romain KUNTZ
kuntz at lsiit.u-strasbg.fr
LSIIT - Networks and Protocols Team
http://clarinet.u-strasbg.fr/~kuntz/




More information about the Support mailing list