Link aggregation at level higher than EtherChannel?

I'm working on a cluster set-up, to which I'd like to create
a fault-tolerant interconnect, _and_ be able to use the aggregate
capacity of the links when there are no failures. As far as I see,
the application to be run on the cluster (Oracle RAC) only allows
for a single interconnect IP address per node.

Below is the "diamond" hardware configuration which would comply
with the fault tolerance requirement, but then the link aggregation
becomes an issue -- plain trunking/bonding/etherchannel won't work
here, as the hosts connect to separate switches.

+-----+
| |
,-| SW1 |-,
+----+ | | | | +----+
| |-' +-----+ '-| |
| N1 | | | N2 |
| |-, +-----+ ,-| |
+----+ | | | | +----+
'-| SW2 |-'
| |
+-----+

Fault tolerance alone would be relatively easy; just arrange for
a local interface failover, and "the network takes care of the rest".

So, any suggestions to aggregate the bandwidth of the two links,
while maintaining fault-tolerance (outage of any single link,
interface or switch)?

Thanks,
--
Wolf a.k.a. Juha Laiho Espoo, Finland
(GC 3.0) GIT d- s+: a C++ ULSH++++$ P++@ L+++ E- W+$@ N++ !K w !O !M V
PS(+) PE Y+ PGP(+) t- 5 !X R !tv b+ !DI D G e+ h---- r+++ y++++
"...cancel my subscription to the resurrection!" (Jim Morrison)

Juha Laiho <(E-Mail Removed)> wrote:
> I'm working on a cluster set-up, to which I'd like to create
> a fault-tolerant interconnect, _and_ be able to use the aggregate
> capacity of the links when there are no failures. As far as I see,
> the application to be run on the cluster (Oracle RAC) only allows
> for a single interconnect IP address per node.

> Below is the "diamond" hardware configuration which would comply
> with the fault tolerance requirement, but then the link aggregation
> becomes an issue -- plain trunking/bonding/etherchannel won't work
> here, as the hosts connect to separate switches.

Unless the trunking/bonding/aggregation software on N1 & N2, and in
the firmware on SW1 and SW2 support "meshing."

> +-----+
> | |
> ,-| SW1 |-,
> +----+ | | | | +----+
> | |-' +-----+ '-| |
> | N1 | | | N2 |
> | |-, +-----+ ,-| |
> +----+ | | | | +----+
> '-| SW2 |-'
> | |
> +-----+

> Fault tolerance alone would be relatively easy; just arrange for
> a local interface failover, and "the network takes care of the rest".

> So, any suggestions to aggregate the bandwidth of the two links,
> while maintaining fault-tolerance (outage of any single link,
> interface or switch)?

Just to be clear - you aren't looking for a _single_ connection to go
faster right, but the aggregate of several connections?

Unless we're talking about 10G here, in broad handwaving terms which
may completely ignore the specifics of your situation, NICs and switch
ports are cheap - how about simply doubling-up and go active-standby
with the switches, two links to each?

rick jones
--
No need to believe in either side, or any side. There is no cause.
There's only yourself. The belief is in your own precision. - Jobert
these opinions are mine, all mine; HP might not want them anyway...
feel free to post, OR email to rick.jones2 in hp.com but NOT BOTH...