We’ve got our costs for this down, which lets us deliver a good price and remain profitable. If you shop around with datacenters – you can get the per-RU cost down to about $50/RU (or less) plus a little extra for bandwidth. (~40 EUR)

We’re hosting this in Dallas, TX where space and bandwidth are both pretty cheap, combining this with using leased equipment – means we dont need to pay off a big upfront cost, so in theory each batch of customers we get are sustainable and profitable from day one. (which is exactly how we like to build a business)

We put a lot of research into selection of where we’re sourcing various parts and hosting – so the result is we can deliver a really competitive service. (We ended up evaluating about ten providers in detail – we had some who were cheaper, but we weren’t too confident about their service in the event something went wrong.)

Re: OSgrid – that’s actually a completely seperate endeavour, and is quite healthy financially (it’s covering costs and putting a little away each month for future purchases). CKrinke could explain more if you asked him; but there’s definetely no dire need for funding there.

If you need money to keep OSgrid online why not ask some sponsoring from Intel and IBM? They get enough in return so I don’t see why they would mind to give you a few 1000 dollars to keep the beta test grid for Opensim online.

We’re not publishing that information publically, because we want to be able to change it down the road as new hardware becomes availible, or provider pricing changes (as it inevitably does).

That being said, if everyone takes the following with a grain of salt (ie, we might change this depending on customer experience, provider pricing, etc) – our standard build out is based around a AMD Phenom 9600 CPU as the base w/ 4GB DDR2, we also have a few Xeon 3220s (”Kentsfield”) with 2GB of RAM for use as stand-by servers (so we can get you setup while waiting for server provisioning). Over the top of this we are using Windows 2008, with the latest versions of .NET installed.

If you are still curious about the specifics, I suggest contacting support as they can probably answer in a bit more detail.

We are artificially limiting the other packages from a quality-of-experience standpoint (since the shared packages do share some resources such as network bandwidth). For dedicated users that limit isnt enforced – since the user isnt sharing any resources. So it’s “not limited” – but it will be the maximum capacity of what the server can deliver.

Under the “standard dedicated” option it shows ” Concurrency: Not Limited ” — what does this mean? Are you able to get more than the 20 standard avatars into a single region? Or is that unlimited for the entire grid?

– Maria