The thinking in this post has come about due to the considerable amount of chat regarding lag, lag compensation, slow hosts, dedicated servers etc etc.
I have searched around and read a few articles on why dedicated servers are unlikely to happen - yes, the cost does seem prohibitive. So the current model, from a cost perspective is the way to go, however, why not expand on the current system?
More years ago than I care to mention, I was an avid player of Doom, Quake and Unreal, granted, they were not on consoles then, but PCs and Macs. The game server engine was built into the actual game, exactly as it is today on consoles. However, in many cases, the developers made the server engine available for download at no cost from their websites - just the server engine, no graphics, nothing, just a pure server. You could set it up, set to automatically run on startup, and run it headless from then on. Hundreds of servers appeared all over the place, run by ISPs, clubs and game enthusiasts.
For this approach, there is near zero development cost (relative to the full game development), as the code has to be written for the full game anyway. So I do not see a barrier from a cost perspective. It is near invisible to the end user, apart from improved gameplay and much reduced lag when they are connected to a server close to them - another win.
Servers all around the world at little cost to A-vision or 3arch!
I know I have simplified certain aspects of this, and it does not address the rather over zealous lag compensation engine built into BLOPS2, but if the implementation of privately run servers took off, 3arch could certainly calm the compensation somewhat.
In the main what do you think of this idea and what is do you think is the barrier to it actually being done?