With the massive push toward cloud computing in the enterprise, there are some considerations that hardware vendors will have to come to terms with in the long run. Unlike the old infrastructure model with hardware bearing the brunt of fault tolerance, the new infrastructure model places all fault tolerance concerns within the software layer itself. I won't say that this is a new concept as Google has been doing exactly this for a very long time (in IT time at least.) This accomplishes many things, but two particular benefits are that load balancing can now be more intelligent overall, and hardware can be reduced to the absolute most commodity parts available to cut cost.
Because the cloud does not need hardware to provide fault tolerance, the hardware required for a 'cloud server' is very basic. I like to think of these servers as netbook equivalents. Bargain bin motherboards, processors, RAM and hard drives can be thrown together to make a low cost commodity cloud server. A 'Cloud OS' and 'Cloud FS' handle the underpinnings as far as operating system and distributed file system. When combined in the right fashion, the Cloud Software Layer along with the underpinning Cloud OS and Cloud FS can literally allow one of these 'cloud servers' to be plugged in and auto-provision itself into the resource pool. When there is a failure of a component or an entire cloud server, the Cloud Software Layer can notify system administrators. Replacement is as simple as unplugging the bad server and plugging in a new one. The server will auto provision itself into the resource pool and it's ready to go. Management and maintenance are simplified greatly.
Looking back at the hardware that will be used to make these cloud servers, last generation surplus parts are perfect for this type of implementation. Each individual server (or node in grid terms) has modest requirements similar to that of a netbook computer. The tasks that these servers will perform are well defined and it is the combination of hundreds or thousands of these servers that provide the real horse power behind the cloud. We see that netbooks can cost as little as $200 and I see no reason why these small cloud servers can not hit the $100 mark as they need no LCD display or peripheral ports, they can use cheaper standard 3.5" hard drives and need no real casing to speak of (depending on the deployment method.) These units can even be racked in shelves of 4 units with direct DC power to each board. There would only need to be a single AC to DC inverter per rack with UPS to ensure power is flowing to the rack as a whole. The amount of heat being created will be far less than with a typical server, and it may even be possible to get the thermal thresholds down to the level where a bare heat-sink (without fans) can be used for the processors. This will also drastically reduce the amount of cooling needed in the data center. The possibilities are literally endless and it gets me excited just to think about this type of stuff.
Of course, all of this is dependent on the intelligence and robust fault tolerance built into the cloud software layer. As I said before, Google has already done this and has been using a similar infrastructure for a long time, so it is not a pipe dream. It is up to the individual hardware vendors such as Sun, HP and Dell to design and deliver a cloud server that will meed the needs of future cloud computing infrastructures. They will also need to deliver it at a cost that reflects the level of commodity the server now represents in the data center.
Oh, one more thing. I just wanted to note that it is not written anywhere that x86 has to be the processor architecture standard for this new breed of cloud servers. I can easily see a custom designed ARM processor fitting the bill.