[support] Re: [Dsmip] NAT Traversal is not working as mentioned in DSMIPv6 draft

[manish Jamwal]

Hi Nassim
Thanks for reply, I will get back to you, after following the steps,
mentioned by you.

Regards
Manish Jamwal

On Wed, Feb 25, 2009 at 3:14 PM, Nassim Kobeissy
<nassim.kobeissy at gmail.com>wrote:

> On the HA side, the BA MUST be encapsulated in UDP. As Kien said, you use
> XFRM. by making a call from ha.c for the "*
> udpencap_encap_out_traffic_start*" method in *xfrm.c* .
>
> You can use it this way:
>
> *udpencap_encap_out_traffic_start(&bce->our_addr,128,
>                                         &bce->peer_addr,128,
>                                         IPPROTO_MH, /*
> path-to-linux/include/linux/in6.h*/
>                                         IP6_MH_TYPE_BACK,
> /*path-to-linux/include/net/mip6.h  */
>                                         &addr4,
>                                         DSMIP_UDP_DPORT,
>                                         (struct in_addr
> *)&bce->nat_info.src,
>                                         (int) bce->nat_info.port,
>                                         MIP6_PRIO_HOME_SIG);*
>
> Take also a look for the *udpencap_encap_out_traffic_end* for deleting
> xfrm states and policies added by *udpencap_encap_out_traffic_start *
>
>
> regards.
> Nassim
> *
>
> * On Wed, Feb 25, 2009 at 10:29 AM, NGUYEN Duc Kien <
> kien.duc-nguyen at univ-paris8.fr> wrote:
>
>>  Hi,
>>
>> *        a. If BU packet is getting UDP encapsulated in IPv4 network,
>> then why not the BA getting encapsulated in UDP. I think both BU and BA,
>> will use the same UDP                     encapsulation/decapsulation code,
>> then why the behavior is not same.*
>>
>>
>> We can use xfrm to encapsulate BU and BA in UDP, by applying the same
>> principle. All the routines necessary are already implemented from
>> Dsmipv6 first release (in src/xfrm.c). You can send email to Nassim for more
>> information.
>>
>> I think that the data encapsulation is more difficult to do, especially
>> for the multiple CoA.
>>
>>
>>
>>
>> *        b. UDP encapsulation is implemented in kernel. So the changes to
>> support encapsulation of BA will require kernel changes also?*
>>
>>
>>
>> No, xfrm on userland will send the requests to the kernel.
>>
>>
>>
>> Regards,
>>
>> Kien.
>>
>>  ------------------------------
>> *From:* manish Jamwal <manish.jamwal at gmail.com>
>> *To:* Sebastien Decugis <sdecugis at hongo.wide.ad.jp>
>> *Cc:* support at ml.nautilus6.org; NGUYEN Duc Kien <
>> kien.duc-nguyen at univ-paris8.fr>; dsmip at ml.nautilus6.org
>> *Sent:* Wednesday, February 25, 2009 9:31:37 AM
>> *Subject:* Re: [Dsmip] NAT Traversal is not working as mentioned in
>> DSMIPv6 draft
>>
>> Thanks Seb for quick response.
>>
>> Yes, I need this feature and willing to look further into daemon code to
>> come with solution. Need some starting pointers from you. Currently I
>> have two question in mind.
>>
>> a. If BU packet is getting UDP encapsulated in IPv4 network, then why not
>> the BA getting encapsulated in UDP. I think both BU and BA, will use the
>> same UDP encapsulation/decapsulation code, then why the behavior is not
>> same.
>>
>> b. UDP encapsulation is implemented in kernel. So the changes to support
>> encapsulation of BA will require kernel changes also?
>>
>> Regards
>> Manish Jamwal
>>
>> On Wed, Feb 25, 2009 at 1:34 PM, Sebastien Decugis <
>> sdecugis at hongo.wide.ad.jp> wrote:
>>
>>> Hello,
>>>
>>> As far as I remember, the implementation is not complete for the NAT
>>> encapsulation part. You should check the code of the daemon, and try to
>>> complete it, if you need this feature. Anyway, it might not be conform
>>> to the latest revisions of the DSMIP specification, either... Some work
>>> is probably needed on the daemon, and you are welcome if you're willing
>>> to provide this work ;)
>>>
>>> Best regards,
>>> Sebastien.
>>>
>>> manish Jamwal a écrit :
>>>  > Hi Kien
>>> >
>>> > NAT Detection seems fine, but the problem is coming in NAT Traversal.
>>> > The problem statement is mentioned below:
>>> >
>>> > If NAT is detected, the binding acknowledgement, and all future
>>> > packets, should then encapsulated in UDP and IPv4. Which is not
>>> > happening in the current mipv6 code, I have applied all the ipv4
>>> > patches ( mentioned in DSMIPv6 mailing list ). Only BU packets are
>>> > getting udp encapsulated, when MN is in IPv4
>>> > network. Please provide some pointer...
>>> >
>>> > Regards
>>> > Manish Jamwal
>>> >
>>> >
>>> >
>>> >
>>> ------------------------------------------------------------------------
>>> >
>>> > _______________________________________________
>>> > Dsmip mailing list
>>> > Dsmip at ml.nautilus6.org
>>> > http://ml.nautilus6.org/mailman/listinfo/dsmip
>>> >
>>>
>>
>>
>> _______________________________________________
>> Dsmip mailing list
>> Dsmip at ml.nautilus6.org
>> http://ml.nautilus6.org/mailman/listinfo/dsmip
>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://ml.nautilus6.org/pipermail/support/attachments/20090225/5025aab9/attachment-0002.htm