Re: Poll: Indicating tunnel capability

Hi Jukka,

On 26.03.2010 05:09 Jukka Manner wrote:
> We need to do a decision on how to indicate that a node supports the 
> tunnel handling mechanism.
> 
> 1. The current specification defines a generic NODE_CAPABILITY object 
> and introduces one flag to indicate tunnel support, T-bit:
> 
>       0                   1                   2                   3
>       0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
>      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>      |A|B|r|r|         Type          |r|r|r|r|        Length         |
>      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>      |T|                           Reserved                          |
>      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
> 
> 
> 2. The other option is to define a 0-length, no flags, object entitled 
> NODE_CAPABILITY_TUNNEL. The presence of this object implies support for 
> tunnels (no T-bit is needed, the whole 32-bit field is reserved).
> 
> So, the question is: do we want to
> 
> 1. keep this generic node capability object and introduce the first flag 
> for this object, the T-bit, or
> 
> 2. introduce an object only for the tunnel capability?

I second Martin's statement and would prefer option 2 as well.

Martin