FWIW, I have been asking the Progress DBA community about their experience with this class of product. Now, admittedly, the focus on Progress may distort the sample, but there are lots of multi-terabyte databases and multi-thousand users in virtually all countries in the world in this sample, so it is not trivial.
The first observation is that, empirically, these people are running into Nutanix in a limited number of sites versus VMWare and proprietary solutions in a significant number of sites. Admittedly, this could be because of the specific technology, but it makes me wonder about the market characterizations.
The second observation is that virtually every site is experiencing performance issues. Almost universally, the decision to use this type of software is imposed from on top as a cost saving measure, without adequate analysis as to whether or not it will really work for all applications. For word processing, e-mail, spreadsheets, and the like, it works great, but these are all small file sizes, low transaction volume applications.
Where it falls down is large database applications with large user counts and high transaction volumes. Performance sucks. With some work it can be improved, maybe, but it is guaranteed to be much worse than putting the database on direct, internal drives … these days, often SSD if performance really matters.
The solution is simple, if one simply recognizes the difference between small file size, low transaction volume applications and large file size, large user base, high transaction volume applications. Just apply the right solution to each problem class … but the bean counters tend to want to impose a universal, “cost saving” solution.