On 1/23/2015 1:33 PM, Brian E Carpenter wrote:
> On 24/01/2015 08:44, Joe Touch wrote:
>> Charlie,
>>
>> On 1/23/2015 11:25 AM, Charlie Perkins wrote:
>
> ...
>>>
>>> We very definitely do *not* want to claim it is an L2 network.
>>
>> Then what is it? A set of L2 networks? Those are interconnected by what L3
>> calls routers.
>
> Not always. They are sometimes connected by what L2 calls bridges or
> switches. In the case of 802.11 they might be connected by intermediate
> wire L2 802.3 networks. The whole thing may then pretend to be a single
> broadcast L2 segment but it isn't, and it doesn't perform like one.
L3 considers all of these a single L2 network. L3 hasn't assumed that
all L2 nets are broadcast for a long time.
> There are many ramifications and some of them are clearer now than they
> were during the PILC study.
The main ramifications are:
- deprecating subnet broadcast (RFC2644)
- assuming that multicast needs IGMP snooping at L2 (RFC4541)
One was known before PILC, one shortly after.
Yes, there's been a lot of work in this area since then too - L2VPNs,
TRILL, and many others.
Again, what's new?
Explaining "hidden terminal" isn't useful; we have RFC3753 (and many
textbooks) for that.
Clearly and specifically highlighting that would be useful, but I
haven't seen that yet.
Joe
_______________________________________________
Int-area mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/int-area
