I'm looking into Always On to see if i can benefit from it, but i struggle to understand if i can or not, and i never previously delt with Always On. So i'm turning to here too see if i can grasp this.

Today, we are running a traditional failover cluster with around 150 databases. It is a collection of alot of diffrent stuff that is in alot of cases hardly used, but we can't throw it away for a million diffrent reasons. But prolonged downtime is unaccepteble (sigh). The environment looks like this:

• No single database is buisness critical (but the amount of DBs makes it critical)
• Databases ranging from 10 MB to 10 GB
• Avg. 30.000 Logical read/s, under 1 Physical read/s
• Low CPU usage (below 20%)
• Two physical servers in two datacenters
• Storage is on two VMAX 10k, one in each center, shared volumes distrubuted with VPLEX, so loosing one center will not affect uptime on storage)
• No need for reporting or anything out of the ordinary
• Most DBs are in Simple mode, with a nightly Full backup, business have no problem accepting the nightly backup as restore point
• A rare few databases are running Full mode and has 30 min transaction backups

The reason this solution was choosen at the time, was licensing and adminstration reason. We wanted to cut back on all the single SQL-servers doing nothing but cost us money. And we have removed about 60 SQL-server installations from the environment.

But now there is talk about how we should deal with this in the future, hardware replacement coming up in a year.

And the problem is, as i stated before, i can't quite get my head around if we can benefit from a change, maybe running AlwaysOn on virtual machines insted?

So my question is: What would you think is the best solution to look at given these factors?