we have a hyper-v 2012 r2 cluster running on a raid volume comprised of 12 4tb sas disks. our controller cannot do pass-through so disks will always show as RAID. we are looking to move to 2016 as soon as its available(on mission-critical VM's)
my question is in 2016 can we add SSD's as a caching tier for this array. basically, create a raid 0 with some SSD's as an upper tier for the Spinning rust underneath. storage speed is fine (nearing 3000Mbps) but Iops are horrible and with enough disk activity from the VM's the whole thing slows to a crawl. hoping adding a SSD tier could help the IOPS situation.
the other thing we have looked at is adding storage spaces direct. we have plenty of 15K sasdrives laying around and each node can take 8 but again with the dell perc 700 we can't do passthrough so storage spaces direct won't let us use the disks as they show as raid even when they are single disks. we would love to fill each node with SSD's and use storage spaces direct but it seems we are blocked from doing this.
eventually, we will be adding an NVME storage server for OS drives and use the other storage server for bulk cold storage. all networking is dual 40Gbe so network speed is not a problem.