Re: [STDS-802-16] mesh net routing
Markus Kiennen wrote:
> macsbug@RESEARCH.ATT.COM wrote:
>> " layer 3 routing can suffer from scalability." is too general a
>> statement. Layer 2 can suffer from scalability as well. All network,
>> especially mesh, can suffer from scalability under some scenarios.
>>
>> One can use layer 3 or 2. Layer 2 can be done with ethCS carrying
>> ethernet MAC address or a few other ways I can imagine depending on the
>> scenarios. 802.11s uses layer 2 but borrows ideas from layer 3. Layer 3
>> is out of scope for 16, but the 16 needs to consider supporting
>> mechanisms.
>>
>> This discussion needs more background and assumptions. Are we talking
>> about the 16-2004 mesh mode, or the new MMR activity in 16j (which has
>> not reached the point to discuss routing yet)? arbitrary topology or
>> constrained, etc. etc..
>
> I'm looking at 16-2004d Mesh Mode with arbitrary topology. If (at a
> bottleneck for example) an IP packet is split into pretty small pieces
> because maybe only one Minislot is available at an intermediate station,
> the receiving station would have to wait with further relaying till the
> whole IP packet is received. With a destination address available in MAC
> it could forward those small packages when received.
>
> Rex Buddenberg already pointed out to me, that 802.3 frames are used and
> those carry the MAC addresses that would allow forwarding without asking
> layer 3 for help. I just didn't and don't know where in the standard
> that is defined. I think i need to have a closer look.
>
> Thanks for your answers!
>
> Markus Kiennen
Hi again,
i was just wondering: As I understand it in 6.3.1.2 the Xmt Node ID
transferred in the Mesh-subheader is the senders ID. I don't see what
this is necessary for. The neighbouring node knows where the data
originates looking at the CID. But if the NodeID would be the
destination ID it would also know where to forward it to.
I hope you can bring light into the darkness of my understanding :)
Thank you.
Markus Kiennen