> * EVE Online reached over 25k concurrent
> users in March, 2006 (no direct link
> available). World of Warcraft hit 200k
> concurrent users

The comparison is one of those apples/oranges deals, because WoW is sharded, and peak concurrency for each shard is considerably less (educated guess: 3,000-4,000 tops), and EVE Online is not a fully embodied 3D world, i.e., graphically it's limited to a spaceship simulation, and therefore way less server-taxing. Which I guess would make this an apples/oranges/bananas deal.

(Not taking anything away from EVE, it's great they've architected it to have such a high concurrency, which is I think the key to its steadily growing subscription number.)

The comparison is highly relevant. These are businesses and technology platforms and must scale to survive and grow. Tony's comment about a 16 acre only supporting a couple of dozen users says that unless Linden Lab is building a Virtual Kansas, they are going to get into deep trouble.

Most people want a Virtual Manhattan.

There are many ways to design these systems - the description of how Second Life is designed should raise a lot of questions. Basically, anything that gets "hot" in Second Life will quickly kill itself. 7000 peak users means these problems rarely occur (think about it, Linden Lab is running nearly 3000 servers to service at most 7000 people!). If Second Life ever grew to the usage rates of Eve, it would be in a whole lot of trouble.

The business comparison of these games is even more valid. Eve has about the same population, a recurring subscription revenue stream from all of its participants, a population that seems to like the game enough to play it on average 3 times as much as SL (I suspect more as peak usage and total hours probably don't match well), and an infrastructure that may have half as many servers (if it is sized like Everquest).

Also, Eve is mostly a Player generated environment, like SL, thus its makers don't need to pump content in to keep it entertaining, unlike most of the other traditional MMOs.

Which do you think will survive?

No one forced Linden Lab to build Second Life the way they did. The statements in the base article that Second Life will scale like Yahoo or Google is flat out wrong - those services can load-balance, SL can't. Google and Yahoo can make reliable predictions on revenue related to CPU and storage and bandwidth, SL cannot - fun in SL is highly correlated with specific chunks of real estate.