If true, thats good going. However, they saved a lot of money by going google-style, using a large cluster of cheapish Debian Linux boxes, with some sort of SAN attached. You can save money this way, which is why I suggested having a look at the clustering and routing bits mentined ^, but there are significant technical hurdles to this (apart from anything else it takes forever to restart a shard).clyne wrote: Though another thing to note that Ryzom was developed within 7 Million or so, was it not?
With some decent I/O added, you could perhaps run a smaller MMO happily on what you have - although you would need to find one that actually ran well on a cluster - writing efficent code for clusters is not especially trivial (this may be one of the reasons Ryzom development has been somewhat sluggish since release).
