Once again, apologies for the slow turnaround time on this. It was a combination of being really busy with my real job along with the poor questions I asked and the way the poll plugin worked. I had to work out how to extract the raw data from the MySQL DB to make some real sense out of it. Also as mentioned, I am pretty sure we had some ballot box stuffing going on so I did some filter based on IP addresses. I am looking around for some better polling software. I likely will have to use something outside of WordPress based on what I have seen so far but hope to put together a better poll that is well suited to multi-question polling.
So this is NOT a scientific poll. I need better software, better questions, and a bigger sample set as well as a method to guarantee unique responses to get something I would call scientific. However it is still very interesting and enlightening, at least to me.
Overall I was pretty surprised to see the penetration of the virtual DCs in production environments. This is not even close to what I have been experiencing out in the field with the hundreds of customers I work with.
So we start off with Lab environments. Good to see so many lab environments, it seems to be the first thing cut in many orgs when they don’t want to spend money thereby turning their production environment into their lab. I shouldn’t have to point out that that can be a bit dangerous. I was surprised there was a response of "No, we don’t need one."; I put that response in as sort of a trick response. The "No, we have a lab domain in the production forest" also kind of surprised me. That being said, I regularly see environments that have no lab environment. It sucks and I usually find out after something has hurt their production environment and I ask how the testing went in the lab and I am fed all sorts of excuses of why they don’t have one and why they don’t need one in the face of an actual outage that would have been exposed had it been tested in a lab environment.
Next up, RODCs. RODCs don’t seem very popular. This aligns with what I generally see out in the world. A lot of folks start with plans for RODCs but then run into the implementation details and decide they don’t want to do it or, for some reason, usually apps, can’t do it.
Of the folks using RODCs, a lot of them are being virtualized. I am absolutely behind that. In fact I have long thought Server Core Virtual RODCs was a really good branch office design.
This is the breakdown of the size of environments of folks that responded. Again I was surprised, this time I was surprised by the number of responses for folks with what I generally consider smaller environments. I am really glad to see it though because it isn’t just the customers with 100,000 users that need to use efficient command line methods to effectively support their environment with minimal costs.
For environments under 25,000 users, virtualization is very heavy in the poll. If you are in this space I could see how people think "everyone" is virtualizing. I just recommend that you really follow the guidelines from MSFT. In smaller environments that I have run into outside of work that have virtualized DCs they usually aren’t following one or more of the guidelines, usually completely or nearly completely compromising their redundancy.
And for the biggest surprise for me was the 25,000 or more users space. This is completely outside of what I see in my real work. It hasn’t penetrated as much as it has in the smaller environments but still, there is a lot of it out there.