Thanks for the speedy turn-around, Suresh.
Lots of snipping to just a few open points…. 3. "It is also fairly clear" Well, that is illuminating :-) Perhaps you want to make statements about the SID elements and not about the clarity of the referenced documents? Sure :-). Suggest OLD: It is also fairly clear that the non-SRv6-SID elements that appear in the SRH SID list are simply IPv6 addresses assigned to local interfaces annd MUST conform to [RFC4291]. NEW: As stated above, the non-SRv6-SID elements that appear in the SRH SID list are simply IPv6 addresses assigned to local interfaces and they need to conform to [RFC4291]. [AF] OK 3. Section 3.1. of [RFC8986] describes the format of an SRv6 SID as composed of three parts LOC:FUNCT:ARG, where a locator (LOC) is encoded in the L most significant bits of the SID, followed by F bits of function (FUNCT) and A bits of arguments (ARG). Would it be helpful to qualify L+F+A = 128 in all cases? Actually not. RFC8986 defines L+F+A <=128 instead and this would be inconsistent with that. [AF] OK, this is important! You have… [RFC8754] defines the Segment List of the SRH as a contiguous array of 128-bit IPv6 addresses …and… Section 3.1. of [RFC8986] describes the format of an SRv6 SID as composed of three parts LOC:FUNCT:ARG 8986 helps me by saying that “When L+F+A is less than 128, then the remaining bits of the SID MUST be zero.” But I think it behoves this document to be unambiguous. “Composed of three parts” does not imply to me “composed of three parts and another part that must be set to zero.” So, perhaps you need a bit more detail and some gentle wordsmithing. For example, Section 3.1. of [RFC8986] describes the format of an SRv6 SID as containing three parts LOC:FUNCT:ARG, where a locator (LOC) is encoded in the L most significant bits of the SID, followed by F bits of function (FUNCT) and A bits of arguments (ARG). If L+F+A < 128, the ARG is followed by enough zero bits to fill the 128 bit SID. 3. When an SRv6 SID occurs in the IPv6 destination address field of an IPv6 header, only the longest match prefix corresponding to the locator is used to forward the packet to the node identified by the Locator. Possibly you mean s/is used/should be used/ Or maybe s/used/used by an SRv6-capable node/ This is written as a statement about what happens today rather than specifying behavior for the node to follow. [AF] Well, I like your confidence about all the implementations that are out there 😊 I think there are two points: 1. You are describing SRv6-capable nodes, I think. 2. You might more accurately provide the normative reference to where this behavior is defined, and then say “As defined in [ref], when an SRv6 SID…” BTW you have “locator” and “Locator” 3. Hence the relevant standard to apply here is [RFC7608] that allows the use of variable length prefixes in forwarding I think 7608 is not a standard. Maybe say specification? But also, I don't think that 7608, as a BCP, "allows" anything. Suggest changing this to Hence the relevant specification to apply here is [RFC7608] that requires implementations to support the use of variable length prefixes in forwarding. Does that work? [AF] Much better 4. A node taking part in this mechanism accomplishes this by using the ARG part [RFC8986] of the Destination address field of the IPv6 header to come up with a new Destination address in some of these flavors. "to come up with" and "flavors" are a bit colloquial. Maybe say "derive" and "mechanisms". Ack on the “derive” part, but “flavor” is a specific term used in [I-D.ietf-spring-srv6-srh-compression] [AF] Hrmph! draft-ietf-spring-srv6-srh-compression is a work in progress, is not a normative reference here, and is not part of this last call. I see it uses “flavor” extensively, but it is a poor choice of word. I tell you what: if the meaning of the word is clear, you can tell me what it means in this context. And if you can tell me what it means, you can use that description in your draft. 4.1. There are a few issues that need to be addressed in the C-SID draft prior to its publication as RFC: Erm, no! You can't have an RFC that chats about the current state of another draft, or that claims it is going to be published as an RFC. Perhaps the best solution is to compress sections 4, 4.1, and 4.2 into a very short note that "Many approaches to SID list compression have been proposed. It is important that any solution preserves the properties of the LOC as described in Section 3." This text was added as requested by one of the spring chairs to specify that the spring document needs to address these issues. It would be great if the 6man/spring chairs and ADs can chime in on this topic. [AF] While we wait for them to speak up, let’s just note that if the intention of this text is to set requirements, you should that in terms of “any specification addressing this problem space needs to provide answers to the following questions.” 5. All of the SRv6 related specifications discussed above are intended to be applicable to a contained SR Domain or between collaborating SR Domains. Hence the behavior of SRv6 SIDs is visible purely within the SR domain and they would be treated solely as IPv6 routing prefixes by nodes that are not SR aware. What is meant by a behavior being visible? Any special behavior associated with SRv6 SIDs are not known or acted upon by non-SR-aware nodes and these nodes use them for forwarding based on the prefix as described in RFC7608. [AF] OK. Perhaps we can rephrase as… All of the SRv6 related specifications discussed above are applicable to a contained SR Domain or between collaborating SR Domains. Nodes either inside or outside the SR Domains that are not SR-aware will not perform any special behavior for an SRv6 SIDs and will treat them solely as IPv6 routing prefixes. I know that the permeability of SR domain boundaries is something that really worries at least one of the current ADs, and it might be good to spend some time discussing what happens when things go wrong and a packet with a SID in the DA field escapes from the domain (this is distinct from the behavior of a non-SR node within the domain). Yes. I certainly do understand that concern and one of the tools in reducing the permeability is moving this traffic to a well known filterable prefix at the borders of the domains depending on the stance of the domain. [AF] This helps so long as we include a remark on what will happen when a packet carrying a DA that uses one of these special prefixes is encountered by a node outside the domain. You possibly need to explain how a non-SR node inside the domain differs from a non-SR node outside the domain. 5. As an added factor of safety, it might be prudent to allocate some address space that explicitly signals that the addresses within that space are not intended to comply with [RFC4291]. As described in "are not intended to comply" means "do not comply"? No. It simply means that compliance to RFC4291 cannot be expected. Are you looking for stronger text for requiring non-compliance? [AF] Then just say “cannot be expected to comply with [RFC4291]” The text “are not intended to comply with” can be confused with “are intended to not comply with” Section 3 above, there is precedent for mechanisms that use IPv6 addresses in a manner different from that specified in [RFC4291]. This would be useful in identifying and potentially filtering packets at the edges of the SR Domains as described in Section 4.1. The SRv6 operational community, which is the first intended user of this block, is requested to come up with conventions and guidelines for the use of this newly allocated address block in line with their requirements. This sounds like you are: - not proposing any specific use - allocating the address space on the off-chance that someone might find a use for it - not suggesting that deployments (or implementations) actually change their current behavior How are you arriving at this conclusion. [AF] I arrive at this conclusion by reading your text! It is “unusual” but not forbidden for an Informational RFC to assign a code point with no use described, but in this case it seems reasonable. What is odd is the paragraph quoted. But… Spring is working on draft-ietf-spring-srv6-srh-compression-02. What address space do you think it can be deployed in? Here are some of the potential options a) RIR allocations b) ULA space c) Something else* (this allocation) I think all of these options have pros and cons and what you think of this prefix allocation might depend on what properties you desire. [AF] So, perhaps you could replace the paragraph as… Future specifications are needed to describe the conventions and guidelines for the use of this newly allocated address block. 6. Obviously, there are many ranges in the registry marked as "Reserved by IETF" and IANA will need help selecting one. Also, since this registry is "IESG Approval" it would be timely to approach the IESG and determine whether they are likely to say "yes" or will need further changes to the document. Those changes should happen while the document is still in the working group. Hmm. Isn’t that what the IESG review process is for? Or are you suggesting an early allocation request prior to advancing the draft so that the IESG can decide if a temporary allocation is worthwhile? If it is neither, can you elaborate on your proposed procedure. [AF] Quite. That is the purpose. However, I’m am just suggesting that you engage your AD now rather than complete WG last call etc. before discovering that the IESG is unlikely to support the proposal. This can be completely informal. I’m not suggesting that you do early allocation or anything like that. I'm surprised that section 7 doesn't point back to the "additional safety" described in section 5. In particular, not using that safety would appear to be a risk. I can certainly duplicate some of the text from section 5 if the WG would find it useful. [AF] Feel free to duplicate text, but I was suggesting highlighting the risk (which is not significantly stressed in section 5), and then pointing back at section 5 as the resolution. Cheers, Adrian
_______________________________________________ spring mailing list spring@ietf.org https://www.ietf.org/mailman/listinfo/spring
