Hi Charlie, hi all,
Am 30.03.2016 um 01:31 schrieb Charlie Perkins
<charles.perkins@xxxxxxxxxxxxx>:
Hello folks,
The list elements of the AODV_INTERFACES conceptual data structure each only
need to have a single element. It is sufficient for the purpose that the
information in that single element should unambiguously identify a single
interface. I think the common name for such a thing is an "interface
identifier", or IID. See RFC 4291.
So, we can say that each element in AODV_INTERFACES is an identifier that is
unique in node-local scope and that allows the AODVv2 implementation to
identify exactly one local network interface. Examples of such identifiers
are the IID as described in [RFC 4291], or a MAC address may be used. By the
way, the definition for IID does not depend on IPv6.
It's about ten times too many words for something intuitively obvious. Oh
well. I hope this at least answers the question.
Regards,
Charlie P.
On 3/28/2016 1:50 PM, Lotte Steenbrink wrote:
Hi Charlie,
Am 28.03.2016 um 19:08 schrieb Charlie Perkins
<charles.perkins@xxxxxxxxxxxxx <mailto:charles.perkins@xxxxxxxxxxxxx>>:
Hello Lotte,
Since it's a conceptual list, for interoperability I don't think it matters
what we specify.
I agree– it’s just that we have to be consistent in how we’re referring to
this list. So we need to write down what exactly is in this list and what
we’re calling that throughout the document. Can you write that down (I think
<https://tools.ietf.org/html/draft-ietf-manet-aodvv2-13#section-4.6>https://tools.ietf.org/html/draft-ietf-manet-aodvv2-13#section-4.6
<https://tools.ietf.org/html/draft-ietf-manet-aodvv2-13#section-4.6> could
make agreat blueprint) until thursday or so? If not I’ll try to come up with
something.
I’d be happy to add an „this is a conceptual list so implement it how you
see fit“ disclaimer though.
But for concreteness we can specify that the conceptual list always has to
contain one interface. If the implementer writes code so that the list may
be actually empty when there's one interface, that's fine with me.
Agree– while we’re at it, I think we can just add an „if there is only one
interface available, the list MAY be left empty“ to resolve Justin’s other
issue.
Regards,
Lotte
Regards,
Charlie P.
On 3/22/2016 3:35 AM, Lotte Steenbrink wrote:
Hi Charlie, hi all,
Am 21.03.2016 um 19:08 schrieb Charlie Perkins
<charles.perkins@xxxxxxxxxxxxx <mailto:charles.perkins@xxxxxxxxxxxxx>>:
Hello folks,
I think the change in terminology is fine, either way.
Okay, change made. :) However, the text you’ve suggested doesn’t specify
what the entries of the set should look like, and if I remember it
correctly, that was the criticism we’ve received. I’m assuming an entry
would contain something like
* Interface.id : the indentifier of the interface configured for use by
AODVv2. Since the format of these identifiers is system-dependent, this
field MAY be extended. This has no impact on interoperability.
But that’s only what I’d need to implement this for RIOT… Do the entries
need to contain anything else?
Also, Justin asked
Also is the list empty if there is only one interface?
– what’s the answer to that? (I’d say the list SHOULD be empty then as it
seems like a waste of space)
Will the new draft be out today?
Nope, I was waiting to hear back from Justin and Christopher– and also the
Security Considerations have some TODOs to resolve, so if anyone has time
to look at that… ;)
Regards,
Lotte
Regards,
Charlie P.
On 3/20/2016 4:12 AM, Victoria Mercieca wrote:
Hi Lotte,
I'd be happy with it becoming InterfaceSet for consistency, would that
be an easy change to make?
Regards,
Vicky
On Sat, Mar 19, 2016 at 10:35 AM, Lotte Steenbrink <
<mailto:lotte.steenbrink@xxxxxxxxxxxx>lotte.steenbrink@xxxxxxxxxxxx
<mailto:lotte.steenbrink@xxxxxxxxxxxx>> wrote:
Hi all,
thinking about it again… Why is the set (or list? charlie’s text has
both words… I’m guessing it’s a set though?) called AODVv2_INTERFACES,
which looks like a constant, instead of InterfaceSet, like the other
sets and lists we have? Is it because it’s administratively configured
and doesn’t change? Or something else? Or is it more or less random?
Regards,
Lotte
Am 17.03.2016 um 11:31 schrieb Charlie Perkins <
<mailto:charles.perkins@xxxxxxxxxxxxx>charles.perkins@xxxxxxxxxxxxx
<mailto:charles.perkins@xxxxxxxxxxxxx>>:
Hello Lotte,
Here's what I sent earlier:
-------- Forwarded Message --------
Subject: Definition for AODVv2_INTERFACES
Date: Fri, 4 Mar 2016 15:38:15 -0800
From: Charlie Perkins
<mailto:charliep51@xxxxxxxxxxxxx><charliep51@xxxxxxxxxxxxx>
<mailto:charliep51@xxxxxxxxxxxxx>
To:
<mailto:aodvv2-discuss@xxxxxxxxxxxxx>aodvv2-discuss@xxxxxxxxxxxxx
<mailto:aodvv2-discuss@xxxxxxxxxxxxx>
<mailto:aodvv2-discuss@xxxxxxxxxxxxx><aodvv2-discuss@xxxxxxxxxxxxx>
<mailto:aodvv2-discuss@xxxxxxxxxxxxx>
Hello folks,
As promised, here is a definition:
AODVv2_INTERFACES: the set of all network interfaces supporting AODVv2
In order to be more pleasing to more people, Section 4.1 should be
revised:
A list of the AODVv2 router's interfaces that support AODVv2 MUST
be configured in the
AODVv2_INTERFACES list.
Regards,
Charlie P.