HA-iSCSI storage nodes

i've been combing net , knowledge bases here find fail-over solution storage not involve san backend.  lay background down development solution i'm messing follows:

4 total physical nodes:

2 hyper-v hosts(ha-clustered failover)-

2 storage hosts(ha-clustered failover)

...so far i've gotten above working in  3 node solution. 2 hyper-v hosts 1 storage node, iscsi target hosting quarem , csv disk.

in config, if lose 1 storage node, lose hyper-v clustering, thought...ok lets add add second storage node fail-over to.

here in lies issue. i've read until points adding not additional storage node, backend server(sas ect...) allows clustering of 2 storage nodes how single storage node above allows clustering of 2 hyper-v hosts.

at point i'm wondering if i'm making sense. or maybe thought process wrong.

in nutshull i'm trying achieve  failover storage nodes without having add backend server cluster 2 storage nodes together.

so far thing achieves 3rd party software means of like:

http://www.starwindsoftware.com/windows-server-clusters

in storage platforms run active-active , continuously syncing each other. there build in mechanism behavior? or required have stand alone san hosting quarem disks , csv disks both hyper-v nodes , storage nodes?

i don't need walk though means, i'd appreciated pointed in right directions, because right i'm going in circles.

no built-in capability.  correct in have found - starwind 1 vendor provides describe. 

a possibility using microsoft have storage shelves shared between storage hosts connected pci raid controllers.  currently, think have seen single vendor has released such raid controller.  raid protecting shared storage on storage hosts, hosts offer either iscsi or smb hyper-v cluster.  assumes 2012.

---

tim

Windows Server  >  High Availability (Clustering)