Hi Zhenqiang,

Many thanks for your detailed review.
Adding 6man WG.

Please see [ZA] in-line.

Thanks

Regards … Zafar

From: li zhenqiang <[email protected]>
Date: Friday, August 23, 2019 at 6:25 AM
To: draft-ali-6man-spring-srv6-oam <[email protected]>, 
spring <[email protected]>
Subject: comments on draft-ali-6man-spring-srv6-oam-03
Resent-From: <[email protected]>
Resent-To: <[email protected]>, <[email protected]>, 
<[email protected]>, <[email protected]>, 
<[email protected]>
Resent-Date: Friday, August 23, 2019 at 6:25 AM

Dear authors and all,

Zhenqiang Li from China Mobile. Please consider the following comments.

1. Section 4.1.2.1, bullet 4, When node N4 receives the packet.
I think it is better to change "The packet gets punted to the ICMPv6 process 
for processing" to "The packet gets time-stamped and punted to the ICMPv6 
process for processing".

[ZA] Agreed. I am making the change in draft-ietf-6man-spring-srv6-oam-01.

2. Section 4.1.2.2, the last bullet.
The sentence "The ICMPv6 process at node N4 checks if its local SID (B:2:C31) 
is locally programmed or not and responds to the ICMPv6 Echo Request" should be 
"The ICMPv6 process at node N4 checks if its local SID (B:4:C52) is locally 
programmed or not and responds to the ICMPv6 Echo Request".

[ZA] Thank for picking the typo. I am fixing it in the latest version.

Since Flags.O=1 is used in this case, N4 should execute the local SID B:4:C52 
after the execution of the pseudo-code specified in section 3.1.1. What is the 
proper behavior of N4 here?  Generate an ICMP parameter problem message to the 
Source with Error code "SR Upper-layer Header Error" according to section 
4.3.1.2 of draft-ietf-6man-segment-routing-header-22?

[ZA] The O-bit is processed before the local SID. Also, you were reading an 
older version of the draft. In draft-ietf-6man-spring-srv6-oam-00 (Aug. 9), 
there is a “Ref2” added in section 3.4. The ICMP parameter problem "SR 
Upper-layer Header Error" is not generated.

To align with the bullets listed in section 4.1.2.1, we'd better add one new 
bullet here with the text "If the target SID is not locally programmed, N4 
responses with the ICMPv6 message (Type:"SRv6 OAM (TBA)", Code: "SID not 
locally implemented (TBA)"); otherwise a success is returned".

[ZA] Addressing in the new revision.

3. Section 4.2.2.1, the last bullet.
I think it is better to change "The packet gets punted to the traceroute 
process for processing" to "The packet gets time-stamped and punted to the 
traceroute process for processing"

[ZA] Addressing in the new revision.

To align with the bullets listed in section 4.1.2.1, we'd better add one new 
bullet here with the text "If the target SID B:4:C52 is locally programmed, 
node N4 responses with the ICMPv6 message (Type: Destination unreachable, Code: 
Port Unreachable). If the target SID B:4:C52 is not a local SID, node N4 
silently drops the traceroute probe".

[ZA] Addressing in the new revision.

4. Section 4.2.2.2, the last bullet.
The sentence "The traceroute process at node N4 checks if its local SID 
(B:2:C31) is locally programmed" should be "The traceroute process at node N4 
checks if its local SID (B:4:C52) is locally programmed".

[ZA] Addressing in the new revision.

Since Flags.O=1 is used in this case, N4 should execute the local SID B:4:C52 
after the execution of the pseudo-code specified in section 3.1.1. What is the 
proper behavior of N4 here?  Generate an ICMP parameter problem message to the 
Source with Error code "SR Upper-layer Header Error" according to section 
4.3.1.2 of draft-ietf-6man-segment-routing-header-22?

[ZA] Same as similar comment above for the ICMP section.

To align with the bullets listed in section 4.1.2.1, we'd better split this 
bullet into two.

[ZA] Addressing in the new revision.

5. Figure 5
The SL for the response from 2001:DB8:1:2:21:: should be 1. The SL for the 
response from 2001:DB8:3:4::41:: should be 0.

[ZA] Good catch! Addressing in the new revision.

6. Section 6.2
Don't we need to define PSP and USP for END.OP and END.OTP?

[ZA] No. This is because END.OP and END.OTP SIDs are always followed by a 
target SID.

My pleasure to work together to move this draft forward.

[ZA] Many thanks and if you are traveling to Singapore, look forward to meeting 
you face-to-face.


Best Regards,
Zhenqiang Li
________________________________
[email protected]
_______________________________________________
spring mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/spring

Reply via email to