Speaking as WG Member and primary author of RFC 8362: 

Hi John, 

On 9/6/22, 1:17 PM, "John Scudder" <[email protected]> wrote:

    Hi Acee,

    > On Sep 5, 2022, at 1:23 PM, Acee Lindem (acee) <[email protected]> wrote:
    > 
    > First of all, I think this IANA OSPFv3 Extended-LSA Sub-TLV 
specification, if it is to be done at all, should be done in a separate draft.

    I guess by “specification” you mean “registry reorganization”, right? I’m 
OK with that, though I think I will wait a little longer to see if other WG 
contributors want to weigh in.

[ACEE] Actually, it is much more than "reorganization" as it is adding lots of 
additional information, i.e., additional "specification". 


    > When we originally created the registry for RFC 8362, the purpose was 
avoid Sub-TLV code point collisions while affording maximum reuse of Sub-TLV 
definitions. It was NOT to avoid reading the specifications in order to 
determine everywhere a Sub-TLV is allowed.

    I agree that reading all the RFCs is (or should be) sufficient. My thoughts 
are as I mentioned in my earlier reply to Ketan — organizing the registry to 
reflect the restrictions imposed by the RFCs serves two possibly-useful 
purposes. First, it gathers the information in one place for quick reference. 
Second and more important, it reduces the chance that the author of a future 
spec might overlook the need to carefully specify where their sub-TLV can and 
can’t be used.

[ACEE] There are a lot of IETF work items that would provide value. However, 
there is only a limited amount of resource and time to complete them. This 
draft has already gone through working group last call and I think this 
additional IANA specification would require more work and reviews than the 
original draft. 

    > In fact, I don’t believe IS-IS had yet adopted this IANA practice when 
this became a WG document in 2007. OSPFv3 Extended LSAs took some time to 
standardize as we held it until there were implementations and there weren’t 
implementations until Segment Routing offered a strong incentive. 
    >  
    > If this information is to be represented in the IANA registry, I’d stay 
away from columns with Y’s and N’s. Note that Sub-TLVs can nested so 
conceivably a given Sub-TLV could be used by other Sub-TLVs.

    Are there existing restrictions on the use of one sub-TLV within other 
sub-TLVs that would need to be captured, or are you just mentioning that for 
completeness? (Completeness is good of course.) If the latter, maybe one option 
could be a catch-all column, called something like “other restrictions on use” 
that can list other restrictions either in line or in a footnote with 
sub-bullets as you suggest.

[ACEE] I don't believe there are any nested Sub-TLVs today but we certainly 
don't want the IANA specification/organization to make adding them unwieldy. 

    > Rather, one could list the Sub-TLVs with the individual usages and 
references (if different from the creation reference) as sub-bullets.

    I can’t comment because I can’t quite picture what you’re describing, and 
how the registry would look?

[ACEE] As it does today except with indented lines as to the usage of the 
Sub-TLVs. The exact formatting would be for the new draft. 

    I guess if we do decide to either abandon the reorganization suggestion 
altogether, or to pursue it as a separate draft, then 
draft-ietf-lsr-ospf-l2bundles should just stick to its existing approach of 
listing restrictions in their own subsections of the main spec, do you agree? 
Recall that we got here (in part) because it seemed strange to me to update the 
registry to list some restrictions, but not all of them.

[ACEE] This would be my choice except I wouldn't add the "L2 Member Bundle 
Attributes" restriction to the IANA registry unless we do it for all the 
Sub-TLVs as you suggest. 

Thanks,
Acee

    Thanks,

    —John

_______________________________________________
Lsr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/lsr

Reply via email to