Re: [Anima-signaling] Draft message on proposed change to GRASP

[Brian E Carpenter <[e-mail]>]

On 23/04/2016 08:29, Brian E Carpenter wrote:
> On 23/04/2016 00:37, Michael Richardson wrote:
>>
>> Brian E Carpenter <[e-mail]> wrote:
>>     > We propose to redefine the locator returned by GRASP discovery accordingly. For
>>     > the IPv6 case that would give:
>>
>>     > locator-option /= [ipv6-locator-option, port-number]
>>     > ipv6-locator-option = bytes .size 16
>>     > port-number = 0..65535
>>
>> This is a good thing.
> 
> OK, I'll wait a day or two for other opinions. Also I think there is
> a case for including the protocol (TCP, UDP or theoretically some other
> transport). If an ASA supports both TCP and UDP it could return two
> discovery responses.

The more I thought about it, the better it seemed to take that approach.
So then I thought about how that would affect message syntax, and this is
the CDDL that I came up with:

locator-option /= [O_IPv4_LOCATOR, ipv4-address, transport-proto, port-number]
ipv4-address = bytes .size 4
locator-option /= [O_IPv6_LOCATOR, ipv6-address, transport-proto, port-number]
ipv6-address = bytes .size 16
locator-option /= [O_FQDN_LOCATOR, text, transport-proto, port-number]
transport-proto = IPPROTO_TCP / IPPROTO_UDP
IPPROTO_TCP = 6
IPPROTO_UDP = 17
port-number = 0..65535
locator-option /= [O_URI_LOCATOR, text]

To allow somewhat regular parsing, I had to include O_IPv4_LOCATOR
and O_IPv6_LOCATOR option codes. I didn't include protocol and port for
URIs because that would be redundant.

Is this implementable? Yes, I changed my code yesterday and it works.
In fact, with this fix, I can run two copies of GRASP in the same
machine and they can negotiate with each other. I ran my ASA Briggs
and ASA Gray modules in one machine and they happily negotiated
for half an hour, until I got bored watching.

(However, I haven't restructured the code to clearly separate the part
that could be in userland from the part that must be in kernel space.)

So, are we good to propose this to the WG?

Regards
    Brian

> 
>> Also consider that the ACP addressing document can provide a multitude
>> (a /96 was suggested) of IPv6 addresses to each node.
>>
>> Is equality of the ipv6-locator-option important anywhere for saying that two
>> conversations involve the same speakers?  or is there another identifier
>> available?
> 
> That's a complicated question. I don't think it affects the current protocol
> design, which discovers and caches addresses per objective anyway. A node could
> contain numerous ASAs and they each could support several Objectives, and if they
> each had a different IP address, GRASP wouldn't care. But it gets back to trust.
> At the moment we're assuming that if a node is trusted, all the ASAs in it are
> equally trusted, so we don't seem to need an additional identifier. If that
> changed, I think we'd have to re-invent the former Device Identity Option in
> GRASP in a slightly different form.
> 
>     Brian
>