Re: daisy chain reposted

Alex Zeffertt <[email protected]>
Newsgroups gmane.linux.drivers.vlan
Organization Cambridge Broadband Ltd.
Message-ID <[email protected]>
On Mon, 5 Dec 2005 15:31:00 +0100
"Tabaro Toni" <[email protected]> wrote:

> 
> I reposted becouse the diagram was incomprensible :-(
> 
> maybe this is out of topic, but:

Probably.

> i have an embedded system with two lan adapters running linux,
> i would like to wire a lot of it in a "daisy chain way":
> 
>          sys1        sys2       sysn-1      sysn
>      +---------+ +---------+ +---------+ +---------+
>      |eth0 eth1| |eth0 eth1| |eth0 eth1| |eth0 eth1|
>      +---------+ +---------+ +---------+ +---------+
>        |	 |     |      |     |     |     |    
> +--------+   +--x--+      +--x--+     +--x--+    
> | switch |    cross-cables
> +--------+
> 
> This approach eliminate the need of expensive switches.

This looks a bit barmy.  Ethernet switches aren't that expensive you
know ;-)


> It's feasible?

Yes.

> how (bridging, routing?)

Bridging.  brctl.

> There is a limit of chaining devices number?
> 


Not as such.  However, for each frame reaching sysn you add load to
sys1, ..., sysn-1.  This wouldn't happen if you just bought a cheapo
switch and used a star topology.

Also, you must bear in mind that the linux bridges use STP. 
(Spanning Tree.)  This will introduce a delay of 30 seconds before the
bridge is up and sys1 can use its IP stack.  The delay will be 60
seconds for sys2, 90 seconds for sys3, and so on.  Even if you disable
STP broadcasts each unit will still disable forwarding for the first
15 seconds while it listens for STP broadcasts.

Finally there is an obvious reliability problem.  if sys4 dies then
you lose contact with sys5,...,sysn.

> Any help will be appreciated.
> 

Alex
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.