Hi Frozen.
Although it is technologically possible to run DotNetNuke on a VPS, I would highly recommend against.
Almost two years ago we were investigating offering virtual servers for DotNetNuke. We spent a lot of time and money setting up the infrastructure and configuring an environment that would be really good for virtualization. However, every single customer who was running on a VPS had performace problems (even though we very lightly loaded our VPS hardware nodes). At the time, we couldn't figure out why everybody was having performance problems (CPU was low, Memory was good, etc) so we ended up having to pull the packages from our website.
With all of the late-breaking news about virtualization, cloud/grid computing, we decided that we wanted to take another look at virtualization solutions for DotNetNuke. We figured that the first time we tried we might have been too early, maybe we set something up wrong, or maybe we just got a "bad egg" so to speak, so we hired Hewlett Packard Consulting to come in and tell us exactly what we needed, the hardware we needed to buy, what virtualization software we should use (HyperV, VMWare, Virtuozzo), how many VPSs per server and all of that. The results, however, weren't what we or HP were expecting.
For one month HP collected CPU, memory, disk IO, bus load, and a lot of other information on 300 servers that were randomly selected from our infrastructure. This included shared hosting servers as well as dedicated servers for customers who gave us permission to collected usage information from their servers. What we learned: If run inside of a virtualization environment, DotNetNuke performs very poorly when compared to either a shared or dedicated solution. We're working on getting permission from HP to publish the study, however, here's the jist of why it doesn't work good:
1) DotNetNuke is a very "chatty" application. It talks to different I/O services (memory, cache, bus, network services) a LOT.
2) When DotNetNuke is "talking" instead of having a small number of "give me LOTS of data" requests, it has an extremely high volume of "give me little bits of data" requests. (This is not just measuring talking to the SQL server, we're talking about bus IO, disk throughput, etc)
3) The relitive cost of virtualizing small IO requests is extremely high compared with the relitive cost of virtualizing large IO requests. This is confusing, but here's an example of what I mean:
Have you ever noticed that when uploading files via FTP, it takes longer to upload 1000-1k files than it does to upload 1-1mb file? That is because each file upload basically has a "header" and a "footer" that get attached to the "here is data" upload command, and when you're working with small files, that header could be just as large (if not larger) than the actual file contents themselves.
After the study was done, we went out to lunch with HP and they pretty much said "Hey, we would love to sell you a $250,000 virtualization solution but it just simply won't work well for your needs."
With all that being said, your DotNetNuke VPS site may run fine for you if you're the only person running DotnetNuke on your hardware node, however if other people on your hardware node are running DNN, you will start to noticeperformance problems.