Hi Zhibo, Please see my reply below.
Thanks, Yingzhen On Tue, Aug 8, 2023 at 6:15 AM Huzhibo <[email protected]> wrote: > Hi Yingzhen： > > Thanks for your comments, please see inline. > > > > Thanks > > Zhibo > > *From:* rtgwg [mailto:[email protected]] *On Behalf Of *Yingzhen Qu > *Sent:* Saturday, August 5, 2023 9:00 AM > *To:* [email protected]; RTGWG < > [email protected]> > *Subject:* Questions about > draft-cheng-rtgwg-srv6-multihome-egress-protection > > > > Hi Authors, > > > > (as individual contributor) > > > > I reviewed the draft, and have the following questions: > > > > 1. > > About PSD definition, in section 3.2, there are the following texts: > > " A penultimate SRv6 Segment Endpoint node is one that, as part of the > > SID processing, copies the last SID from the SRH into the IPv6 > > Destination Address and decrements the Segments Left value from one > > to zero. > > " > > " The PSD operation *only takes place at the egress node *and does not > > happen at any transit node. When a SID of PSD flavor is processed at > > a transit node, the PSD behavior is not performed as described in > > the pseudocode below since Segments Left would not be 1 or 0. > > " > > so to my understanding, P2 is the penultimate endpoint in the example in > Figure 1, correct? however there is "PSD-flavored SID" on PE3. Please > clarify. > > [Zhibo] Yes, your understanding is correct. We will modify the description > in the next version to avoid ambiguity. > > > [Yingzhen]: Please do. The second quoted text here is not correct. 2. > > In Section 4, "After the configuration, PE1 determines that PE3's backup > SID is PE4's VPN SID through the routing optimization strategy of BGP." . > Can you please elaborate how PE1 decides PE4 to be the backup SID? What if > the path for P2 to reach PE4 is P2->PE3->PE4? > > [Zhibo] According to the BGP route selection principle, the ingress PE > node selects the preferred route as the primary node and the second-best > route as the backup node. > > After egress protection is enabled, if PE3 fails, P2 forwards traffic to > PE4. If PE3 is the next hop of P2, P2 will forwards traffic to PE4 through > the TI-LFA backup path. This solution does not affect the implementation of > this solution. > > > [Yingzhen]: More details about this should be added to the draft, such as topology limitation, how ti-lfa may be used. 3. > > In section 3.4, it mentions that P2 could be several hops away from the > egress node. In this case, if PE3 fails, how does P2 detect the failure of > PE3 quickly? What if a link between P2 and PE3 fails? > > [Zhibo] > > P2 detect the failure of PE3 via: > > • IGP convergence > > • BFD detection > > > > Traditional FRR also maximizes the protection scope and preferentially > considers node faults without distinguishing whether a link fault or node > fault occurs. > [Yingzhen]: P2 and PE3 might be in different ASes, then IGP convergence may not be helpful in this case. Please add some explanations in the draft. > > > Thanks, > > Yingzhen > > > > >
_______________________________________________ rtgwg mailing list [email protected] https://www.ietf.org/mailman/listinfo/rtgwg
