vfs.dev.read and vfs.dev.write were recently fixed to allow more flexible device names to be used as an argument. Something similar needs to be done for Solaris. As it stands now, the only device names it accepts are things like "md#", "sd#", "ssd#". The problem is that sometimes these mappings can change during a reboot, especially the 'ssd#' devices when the underlying device is something like an iSCSI volume.

At the very least, the argument should accept the disk device [ie. "c0t0d0", "c0t1d0s1", "c5t500000E0125A7270d0"].
Even better would be the ability to use a zfs device or zpool as an argument as well [ie. "rpool", or "tank/data" ], or even to use a mounted path [/home, /var]

I don't know how the server actually pulls this data, perhaps it uses iostat? By device the device names in iostat are as currently required [md10, sd1, ssd4]. It is possible to map the sd# to the actual disk device in Solaris:

paste -d= <(iostat -x | awk '{print $1}') <(iostat -xn | awk '{print $NF}') | tail +3

Which just gives a list like this:

md40=md/d40
md41=md/d41
md42=md/d42
sd0=c1t0d0
sd1=c1t1d0
sd3=c0t0d0

Hopefully that will be helpful.