| Hi: While I am not a SRv6 expert, I think I understand the essentials. And I am writing this with dhc wg co-chair hat off. In terms of DHCPv6 aspects, I do have two issues to point out: 1. In section 4.2, “When the sum LB-len, LN-len, Fun-len, and Arg-len exceeds 128 bits, the packet is deemed erroneous and must be discarded.” This kind of goes against the norm for DHCPv6. And is different for T1 greater than T2 and preferred vs valid lifetime handling in the document. Instead, this should just invalidate that particular option and processing should continue. A “minor” configuration or implementation issue should not prevent other DHCPv6 data in the packet from not being used and can cause a packet storm as clients keep sending messages without making (at least some) progress. 2. For Prefix Delegation, which seems to be the model for SRv6 locators, Confirm messages are not used. Hence, it should be removed from section 5.2? And a couple of minor nits … 1. In section 4.2 there is a reference to 8415bis section 10.2.10.1. The “The details about use of these values…” is a bit underclear as to whether it is referring to infinity values (which are never mentioned in 10.2.10.1) or just the lifetime values in general. Perhaps replace “these” with “lifetime”? 2. In section 5.1, first paragraph it might be better to use Solicit, Request, Advertise, and Reply instead of lowercase? Thanks. - Bernie Volz On Aug 28, 2025, at 4:22 PM, Alvaro Retana <[email protected]> wrote:
|
_______________________________________________ spring mailing list -- [email protected] To unsubscribe send an email to [email protected]
