| The reference to "no physical ports" means that routing in the
VNswitch always uses what we call a "routing over bridging" model.
What this means is that you collect sets of bridge ports together
(using VSDs) and treat them as a single interface for routing. So,
for example, you might create 2 VSDs: A and B. In A you might place
ports 1 and 2 and in B you might place ports 3 and 4. You can now
create a routing virtual interface for each of the VSDs A and B.
On those two virtual interfaces you can configure IP routing just
like you would have done on a real interface (you can assign
one or more subnet addresses, turn RIP on or off, turn OSPF on
or off, etc.).
So, the short answer to your question:
>Now , if I have a stand Alone VNswitch 900 and want to ROUTE between front
>ports, let say 3 and 4 on different IP subnets. Will this be supported with
>the IP routing code due soon ????
is "yes", but you will need to decide how to configure ports 3 and 4
into VSDs. You could place them in separate VSDs if you do not want
any bridging connectivity between those ports. Or, if you do want
bridging connectivity between them then you can put them in the same
VSD and configure the virtual interface for that VSD with both subnets.
In that case packets between the two IP subnets will be routed by
the VNswitch but within the same VSD. We have a further trick for
this case which we call "ARP routing" which can be used to avoid having
to always route the packets. If you configure ARP routing (which requires
both VNswitch and IP host configuration) then you can create a bridging
path between systems in different IP subnets - this can give higher
aggregate performance since the VNswitchhas higher bridge forwarding
performance than routing performance.
Chris
|