Re: [Detnet-dp-dt] An optional scheme that can support Detnet with a single layer of PW label

[Jiangyuanlong <[e-mail]>]

Loa, please see my comments in line.

Thanks,
Yuanlong

-----Original Message-----
From: Detnet-dp-dt [mailto:[e-mail]] On Behalf Of Loa Andersson
Sent: Wednesday, February 22, 2017 11:54 AM
To: Jiangyuanlong; Jouni Korhonen
Cc: [e-mail]
Subject: Re: [Detnet-dp-dt] An optional scheme that can support Detnet with a single layer of PW label

Yuanlong,

Very interesting, and I think we need to spend some time on this. I think you are right bringing up the ingress and egress single point of failure. The drawback is that the CE's gets more complicated.
[YJ] Thanks;)

This is still MPLS, but I miss the label stacks to understand the processing.

And I have one specific question on slide 8. CE1 sends two identical packets to PE1 and PE2, each PE asdd a sequence number, right?
After that PE1 sends it packet to the out port for PW1 and D-PW1 and
PE2 send its packet to the out port for PW2 and D-PW2.

Before the packets are send out on PW1 and PW2, filtering (elimination) takes place. The sequence number is used for that? How do we make sure that the sequence numbers generated by PE1 and PE2 are the same?
[YJ] Yes. If CE is 802.1CB compatible, we can use its sequence number directly. Otherwise, a Sequence Generator is needed on a PE attached with AC, how to generate the same seq on two PEs may pose a challenge. Maybe a hash module which takes a whole frame as an input can merely work.

Also the filtering functions on PE1 will handle a pair of packets intended for PW1. One of the packets come from the AC in port and the other from the P-PW1 in port, what is used coordinate the packets?
[YJ] there is no coordination, but the encapsulation will be the same (all in Ethernet for Ethernet frame).

/Loa

On 2017-02-21 17:30, Jiangyuanlong wrote:
> Hi all,
>
> The intention of this slide is to provide an optional scheme that can support Detnet with a single layer of PW label.
> For the time being the slide demonstrates that it is possible in the data plane.
> Your comments are very welcome.
>
> Best regards,
> Yuanlong
>
>
> -----Original Message-----
> From: Detnet-dp-dt [mailto:[e-mail]] On Behalf Of 
> Jouni Korhonen
> Sent: Sunday, February 19, 2017 3:29 PM
> To: Loa Andersson
> Cc: [e-mail]
> Subject: Re: [Detnet-dp-dt] FRER example.. maybe a way forward? was 
> Re: quick notes from call 2/14/15
>
> Loa,
>
>> On 18 Feb 2017, at 23:26, Loa Andersson <[e-mail]> wrote:
>>
>>
>>
>> On 2017-02-19 15:22, Jouni Korhonen wrote:
>>>
>> <snip>
>>>> I don't follow - let us say we have one Ethernet and one TDM PW 
>>>> crossing the domain, id e use the same d-pw label, how do we know 
>>>> which one we do Ethernet NSP and which one we do TDM NSP on??
>>>
>>> Of course each e2e detnet flow have their own d-pw. Otherwise you cannot make a difference between them. Being the same means an e2e detnet flow would have a single d-pw label throughout the detnet/administrative domain in each x-PE.
>>>
>> OK . so what you are saying is that replicated packets will have the same d-pw? I agree to that!
>
> Yes.
>
> - JOuni
>
>>
>> /Loa
>>> - Jouni
>> <snip>
>> --
>>
>>
>> Loa Andersson                        email: [e-mail]
>> Senior MPLS Expert                          [e-mail]
>> Huawei Technologies (consultant)     phone: [phone]
>
> _______________________________________________
> Detnet-dp-dt mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/detnet-dp-dt
>

-- 


Loa Andersson                        email: [e-mail]
Senior MPLS Expert                          [e-mail]
Huawei Technologies (consultant)     phone: [phone]

_______________________________________________
Detnet-dp-dt mailing list
[e-mail]
https://www.ietf.org/mailman/listinfo/detnet-dp-dt