3560 isl and dot1q

The integration seems alot tighter to me than a trunk from one side of the box to the other. You also state that its 802.1d, the IEEE Bridging protocol, and then reference ISL, the cisco proprietary trunking protocol???

So, I guess I don't understand your concern, or what you think may be going on. A 3560 doesn't behave at all like connecting a layer-2 switch up to a cisco router. It does behave alot like a layer-3 switch through and through.

Hi,

Both dot1q and ISL are supported by this Catalyst model when trunking with another switch. You change the frame encapsulation with something like switchport trunk encapsulation in interface config

Regards, Gabriele

"a" ha scritto nel messaggio news:bGBzh.932233$1T2.803622@pd7urf2no...

Thanks for your reply. Lets break down my question. First of all, what is the difference between layer-3 switching than the layer-2 switching?

In my opinion, the 3560 is trying to simulate the layer 2 switching, which is a switch connected to a router, in one box. The trunk between the router and the switch can be configured as ISL or dot1q. Is it right?

Thanks

layer-2 switching keeps all traffic in a VLAN isolated on the bridge layer. layer-3 switching transitions the layer-2 traffic onto an IP layer. There doesn't have to be a .1q trunk involved in order for this to happen.

I'd assume, internally the 3560 has some sort of flags associated with each path going through it, with some huge crossbar hardware to route the paths appropriately. There's probably some hardware slots devoted to layer-3 transitions, such that with a packet comes along that path, it'll get punted into the CPU to put it out onto the layer-3, and do whatever layer-3 stuff that is needed. This is probably one reason (of a few) that the smaller cisco switches generally can do so few VLANs.

Do you have much experience with the old products using the MFSC cards?

Why would you assume that? Cisco's layer-2 and layer-3 switches look almost identical to configure. Cisco's layer-3 switches infact, start out to be identical to their layer-2 switch for a default configuration, if you do nothing else, you'll have a basic layer-2 switch with the 3560 out of the box.

I'd assume that the transition from layer-2 to layer-3 is much lower, and much more tightly integrated than a trunk from a switch over to a router. (like you might have in some of Cisco's old kit, like the 4000 over to the MFSC, or the older 5500 over to the MFSC, but these are old long EOL'd hardware replaced by better hardware).

My testing of cisco layer-3 switches vs. a layer-2 switch trunked to a layer-3 router (ie. like the MFSC based setup) show great differences in the way things like heavy traffic loads behave. I'd assume that it happens much lower on the hardware level than up on the virtual level like you are categorizing these boxes.