Re: 答复: [spring] 答复: Genart last call review of draft-ietf-spring-ipv6-use-cases-10

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Xiaohu,

Mapping can be also well known .. for example from the RFC clearly defining the functions. That clearly removes the need of *any* control plane to dynamically carry the values. 

Cheers,
R.

On Tue, May 9, 2017 at 7:06 AM, Xuxiaohu <xuxiaohu@xxxxxxxxxx> wrote:

Nobody said that such mapping must be distributed in a routing protocolJ   We just said that such mapping is always required no matter we use an MPLS label as function instruction or using the latter part of an IPv6 address as function instruction.

 

Best regards,

Xiaohu

 

 

发件人: rraszuk@xxxxxxxxx [mailto:rraszuk@xxxxxxxxx] 代表 Robert Raszuk
发送时间: 201759 12:21
收件人: Xuxiaohu
抄送: stefano previdi; Stewart Bryant; Alvaro Retana (aretana); draft-ietf-spring-ipv6-use-cases.all@xxxxxxxx; gen-art@xxxxxxxx; IETF Discussion; spring@xxxxxxxx
主题: Re: [spring] 答复: Genart last call review of draft-ietf-spring-ipv6-use-cases-10

 

Hi,

 

Can you point me to any SRv6 document mandating that such mapping MUST be done in a routing protocol ?

 

Thx,

R.

 

 

On May 9, 2017 05:16, "Xuxiaohu" <xuxiaohu@xxxxxxxxxx> wrote:

Hi Robert,

 

Except the function indicated by all zeros, it does require mappings between functions, function indications and locators. Take tenant network function as an exampleJ

 

Best regards,

Xiaohu

 

发件人: spring [mailto:spring-bounces@ietf.org] 代表 Robert Raszuk
发送时间: 201758 20:13
收件人: Stewart Bryant
抄送: spring@xxxxxxxx; IETF Discussion; gen-art@xxxxxxxx; draft-ietf-spring-ipv6-use-cases.all@xxxxxxxx; stefano previdi; Alvaro Retana (aretana)
主题: Re: [spring] Genart last call review of draft-ietf-spring-ipv6-use-cases-10

 

 

Hi Stewart,

 

> ​

Yes, but those are required anyway to support SRv6

 

That is incorrect. 

 

Please watch Les's presentation during last IETF where he clearly and correctly stated that IGP extensions for SRv6 are optional and not required (unlike in the case of SR-MPLS where such signalling of global labels is indeed necessary to build "global labels" based forwarding). 

 

As a matter of fact I knew this will generate confusion in IETF and recommended to make it very clear in the drafts. Distribution of SID functions does not need to be carried in routing protocols. And SRv6 SID locator function is native to IPv6 routing hence no extensions needed. Sure they can be carried in IGPs or BGP but this is just an option not necessity. 

 

Best,

Robert.

 

 

 

On Mon, May 8, 2017 at 12:15 PM, Stewart Bryant <stewart.bryant@xxxxxxxxx> wrote:



On 05/05/2017 11:17, Robert Raszuk wrote:

And to add one observation ..

Stewart makes a point that SR-MPLS can be deployed without mpls control plane.

Well it sure does not require LDP however IGP or BGP extensions for SR-MPLS signalling is also an example of mpls control plane ... even if much simpler than traditional cases it is still required.

 

​​

Yes, but those are required anyway to support SRv6. You always have to provide the mapping between the function, the function identifier and the function location.

- Stewart

 


_______________________________________________
spring mailing list
spring@xxxxxxxx
https://www.ietf.org/mailman/listinfo/spring

 



[Index of Archives]     [IETF Annoucements]     [IETF]     [IP Storage]     [Yosemite News]     [Linux SCTP]     [Linux Newbies]     [Fedora Users]