[j-nsp] NG-MVPN RPT-SPT mode no receiver CE limitation?

Krasimir Avramski krasi at smartcom.bg
Wed May 28 02:50:09 EDT 2014


Hello Vladi,

I didn't mention you are interested in RPT-SPT mode - I just thought you
have local PE receivers issue that was typical in early days of mvpn and of
course we used  SPT-only with SSM mode.There were inherent pfe
infrastructure problems with composite nh needed to dealt with direct ifls
and p2mp (especially on source PE).

Best Regards,
Krasi


On 27 May 2014 17:42, Vladislav A. VASILEV <vladislavavasilev at gmail.com>wrote:

> This is a snippet from "Deploying MBGP Multicast VPNs":
>
>
> *At the time of this writing, Junos OS does not support RPT-SPT mode if
> the CSources*
>
> *or C-Receivers are locally connected to the PEs. It is mandatory to have
> a*
>
> *CE between a PE and a host. This limitation is specific of RPT-SPT, it
> does not affect*
>
> *the default SPT-only mode.*
>
>
> The book is from May 2011, so I assume things must have changed. Otherwise
> what I have should not work at all. I guess I'll wait for JTAC to get back
> to me.
>
>
> Thanks,
>
> Vladi
>
>
> On Tue, May 27, 2014 at 3:27 PM, Krasimir Avramski <krasi at smartcom.bg>wrote:
>
>> I have had customer using exactly the same setup (in production) without
>> any problems on receiver PE (static igmp local receivers for faster channel
>> zapping). Now I can't remember if igmp was effectively filtered on
>> downstream access equipment. The problem with local receivers was only on
>> Source PE ( if you have type-7 C-M route then Source PE local receivers
>> would stop receive mcast traffic - 9.6 release, DPC hardware). The work
>> around was to deploy another vrf on source PE to terminate local receivers
>> and source mcast interface and mvpn vrf serving only remote PEs). On 11.2
>> it was dependent on type of hardware used(MPC, DPC) for interfaces
>> connecting multicast source, mpls core and local receivers - for example if
>> DPC hardware was used for ifls connecting mcast source, mpls core and local
>> receivers then it was OK - every other combination (including only MPC
>> hardware) affected local source PE receivers or remote PEs not receiving
>> multicast traffic at all.
>> I really don't know current state with latest software/hardware, but
>> rather you are hitting some bug.
>>
>> Best Regards,
>> Krasi
>>
>>
>>  On 27 May 2014 14:13, Vladislav A. VASILEV <vladislavavasilev at gmail.com>wrote:
>>
>>>  Does anybody know if this limitation still exist:
>>>
>>> Source CE --- Source PE --- MPLS Network --- Receiver PE --- Static IGMP
>>> under VRF interface (no Receiver CE)
>>>
>>> It works but the problem is that for no reason multicast traffic would
>>> stop
>>> flowing (within a few days) due to BGP withdrawing all route types
>>> (except
>>> neighbour discovery). Deactivating/activating the routing instance under
>>> the receiver PE solves the problem for a few days and it would then
>>> happen
>>> again. I've had a JTAC case opened for a few weeks now, but I still have
>>> not been given a clear answer as to whether or not this is supported.
>>>
>>> Thanks,
>>> Vladi
>>> _______________________________________________
>>> juniper-nsp mailing list juniper-nsp at puck.nether.net
>>> https://puck.nether.net/mailman/listinfo/juniper-nsp
>>>
>>
>>
>


More information about the juniper-nsp mailing list