Hi there,
we are going to revisit our network. Currently we have
core1: 5406
core2: 2910al
(connected with 10GbE)
At the moment they "only" serve our virtualization environment via balance-tlb (one host is connected to both switches) for xen-traffic, data and iSCSI.
All downlinks to the edge switches are only connected to core1 (via LACP) at the moment.
On my recherche through the web I stumbled accross distributed trunking which seems to be the jack of all trades for our virt-cluster (I know that the 2910 is not capable of DT, but we would go for a new redundancy-core then).
But I stumbled about some downsides and on some contradictory announcements:
* A lot of sources state that IP routing is not possible with DT enabled - other source state that this should be possible with never firmware
* official 5400 page on HP ProCurve ( 1. ) state that DT is possible to servers AND switches - all other sources I found state that only connection to servers is possible
* All source I found state that IGMP is not supported with DT
Ideal scenario would be to use DT-LACP for the servers and the downlinks to the edge switches to achieve redundancy and the full LACP performance. But we need routing and IGMP support.
Any ideas and suggestions? Any answers to these three points?
Thanks a lot in advance,
phix
ressources:
1.) http://h17007.www1.hp.com/us/en/networking/products/switches/HP_5400_zl_Switch_Series/index.aspx#tab=TAB2