why we need 'bridge'

Hi,

many layer2 ethernet equipment vendor, Cisco as example, have a concept (I don't really know if it is correct to call it concept?) of a 'bridge', i.e. on cisco CLI one needs to explicitly create bridge:

#conf t #(config) bridge 1 protocol ieee #interface fe0 #bridge-group 1 #interface fe1 #bridge-group 1 ...

Why is it necessary ? As long as it is a switch, all the ports by default are in switchable, or this allows to have a number of independent bridges within one single physical device? Thanks in advance.

Mark

Reply to
Mark
Loading thread data ...

..

This sounds suspiciously like homework already, so you only get a hint.

Its unlikely that somebody setting up a cisco would bridge two ethers together, but they support other kinds..

Reply to
Doug McIntyre

Thanks for reply. It's not a homework, although I'm student :) My guess after some research was that cisco originally supported multiple technologies (for ex. token ring, ethernet, fddi etc.), and it was naturally to allow a customer to bridge these ports in a single bridging entity.

But I wasn't sure if my thoughts were right or wrong. Certainly I searched cisco.com, but they don't give much theeoretical background in this item though.

Mark

Reply to
Mark

Indeed, WAN bridges was the primary use for bridge groups, not necessarily LAN to LAN bridging..

Reply to
Doug McIntyre

We have the concept of L2 ports (aka "switchports") and L3 ports ("routed interfaces".) The former ports naturally bridge (forward at layer 2) amongst themselves if they are configured to be in the same VLAN. The latter port flavor naturally route (forward at layer 3), but can be made to bridge if put into the same bridge-group as shown above.

In olden times, back when a variety of L3 protocols was prevalent, a Cisco router interface could be configured to route some subset of protocols (by configuring it with an L3 address in the given protocol), and then bridge other protocols (by putting the interface into a bridge-group.) Nowadays, with everyone running IPv4 everywhere, you use bridge-group to extend an IPv4 broadcast domain ("subnet"), and you use routed interfaces to delimit the broadcast domain.

The prevalence of switchports is the distinguishing mark of a device that we call a "switch"; the prevalence of routed interfaces is what causes us to call a device a "router". Although nowadays a device like the 881 has 4 switchports (Fa0..3) but only 1 routed port (Fa4), but is nonethleless called a "router" rather than a "switch" - go figure.

I seem to recall that _Interconnections_ by Radia Perlman had a useful discussion of this sort of stuff.

Aaron

Reply to
Aaron Leonard

there are even now some networks that need bridging and a router would not have all interfaces on the same switch module for hardware L2 switching support - protocols like LAT have not died completely and bridging OSI CLNS was common at 1 point in carrier networks for managing SDH and similar kit

This is usually on a router between Ethernet LANs and WAN interfaces.

More recently Ethernet is taking over the high bandwidth WAN role, but even there routers get used on the LAN to WAN border to allow QoS + rate limiting etc in ways switches struggle with.

A network i used to work on used bridging between the loopback and ATM PVCs for management traffic (i think - this is a while back)

finally

Reply to
Stephen

Put multiple switch modules such as the HWIC-4ESWs into a Cisco router and the ports within the module switch L2 locally, but the modules are not connected at layer 2......

Still got lots of OSI CLNS running for SDH and other management = note carrier networks will be happily running kit well past the sell by dates.

Reply to
Stephen

Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.