Re: encoding Subnet ID in link-local addrs - problem statement

[Alexandre Petrescu <[e-mail]>]

Le 25/04/2019 à 12:10, Ole Troan a écrit :
> Hi Alexandre,
> 
>> These are three tries for a problem statement:
>> 
>> 1 A rejected Errata to RFC4291 "IPv6 Addr Archi" on the topic of 
>> link- local addresses 'would need' a draft.  (The errata ID is 
>> 4406; the URL is https://www.rfc-editor.org/errata/eid4406)
>> 
>> Do you disagree with the recommendation of this Errata?
> 
> The errata claims an inconsistency. I do agree with the errata 
> rejection if that's what you are asking.

It is not that that I am asking.

I am asking this: do you disagree with the Verifier Notes text? (I am 
asking because a part of the Verifier Notes seems to me to be a 
recommendation.)

For your convenience, this is the part from the Verifier Notes:
> [...] The re-definition of the link-local address would need to be
> driven by a draft updating RFC 4291.



> Whichever way I try to interpret this "try #1 for a problem 
> statement", I cannot find a reading that resolves into an actual 
> problem.
> 
>> 2 IPv6 link-local addresses are typically self-configured according
>> to 4 RFCs and relying on the fe80::/10 IANA allocation, RFC 54 0
>> bits, and RFC MAC-based 64bit Interface IDs.  In some cases, it is
>> advantageous to manually configure link-local addresses.  This is
>> useful for easy remembering by humans, and for parameter resilience
>> during network interface replacement.
>> 
>> Manual configuration of an LL may use short IID and Subnet ID The 
>> Subnet ID presence in the link-local address is useful in some 
>> wireless settings where the subnet structure parameters depend on 
>> the link locality.  Other settings may also benefit.
>> 
>> When manually setting the link-local address it is necessary to 
>> know the length of the prefix of the subnet on which this 
>> link-local address is present.  This length is necessary for 
>> on-link determination.
> 
> There is nothing in our documents that prohibit you from doing manual
> configuration of link-local addresses.

I agree.

But some OSs prohibits it.  The reasoning behind that OS prohibition is 
invoking this RFC.  It is an interpretation of this RFC that leads to 
prohibit people on that OS to do manual configuration.

> You can achieve easy to type
> and addresses you can remember, and that are different per-subnet
> simply by doing something like:
> 
> FE80::<16-bit subnet id>:<48-bit "station-id">

YEs, fe80::1:1 is a solution too.  But others suggested other solutions 
like fe80:1::1.

The fact that there are two distinct solutions makes it an even stronger 
problem.

On my side, I would not agree with
FE80::<16-bit subnet id>:<48-bit "station-id">
because these ids are part of an Interface ID.

Subnet ID and Interface ID are distinct parts in an IPv6 address, as 
defined by RFC4291.  Can not put a Subnet ID in an Interface ID.

> 
>> Implementation status: manually configuring link-local addresses is
>> an operation permitted by some Operating Systems but forbidden by
>> others. The ones that permit it need a parameter for the length of
>> the prefix. The length of the prefix is sometimes 10 (cf. IANA 
>> allocation), at times 64 (cf. RFC assignment), at times 128 (cf. 2
>>  RFCs) and other intermediary values are known to work.
>> 
>> Is this 2nd try along your line of thought?
> 
> Sure,

Thank you.

If you you do not mind, I will tweak this text a little, and add it in 
the draft about the prefix len of LL addresses.  I will put it as a 
separate section.

> and what with my solution doesn't solve your problem?

It does solve the problem, but in a slightly different way than 
fe80:1::1 does, and which seems little compatible with RFC 4291 Subnet 
ID - IID distinction.

>> 3 more problem statements about 'some wireless settings' can be 
>> found in the IPWAVE WG Problem Statement draft.
>> 
>> Is this 3rd kind of problem statement something that may be 
>> answering your request to provide a problem statement?
> 
> Sorry, I see nothing in draft-ietf-ipwave-vehicular-networking that 
> directly points to the length of the link-local prefix being a 
> problem.

The problems in that draft are probably of less importance to 6MAN WG 
specialised reader.  Your reaction makes me think it is not this kind of 
3rd problem that you requested when you asked for a Problem Statement.

Probably number 2 is, and less probably number 1 is.

Alex

> 
> Cheers, Ole
>