[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [DNA] L2 link and L3 link
> If several Ethernet segments are bridged together, they form a link. If a host
> moves from one Ethernet segment to the another, it hardly notice a change.
The host will notice "link down" and "link up" events. It may also notice
that it is required to do another 802.1X authentication, possibly with a
different device as identified in the EAP-Request/Identity exchange.
> But if 802.11 APs and an Ethernet segment are bridged to form a link, it's
> more complex. Assume there are two APs and a router on an Ethernet
> segment as below.
>
> +----+
> | AR |
> +----+
> |
> +-----+--------+
> | |
> +----+ +----+
> |AP 1| |AP 2|
> +----+ +----+
>
> +----+ +----+
> | H 1| |H 2 |
> +----+ +----+
>
> Each AP forms a different BSS. And if a host moves from AP1 to AP2, it will
> detect distinct link-layer change.
This case is not really different from the Ethernet case. "Link down" and
"Link up" events are observed, along with L2 authentication.
> Also assume host 1 is attached to AP1 and host 2 is attached to AP2. I think
> it's reasonable to say that they are at the different L2 link. But they are at the
> same (L3) link.
Both hosts are within the same broadcast domain, so I'm not clear why you
say that the L2 link is different. Or do you mean that the "point of
attachment" is different? If so, then both cases exhibit a change in the
point of attachment.
> We'd better clarify this lest there should be unnecessary confusion and arguments.
I'd suggest that the best way to eliminate confusion and unnecessary
arguments is to use existing, standard terminology. Creating new
definitions of words like "link" or "subnet" will not help.