ARM servers represent an important shift in server-based computing. A traditional x86-class server with 12, 16, 24 or more cores delivers increased performance by scaling up the speed and sophistication of each processor, using brute force speed and power to handle demanding computing workloads. By comparison, an ARM server employs many smaller, less sophisticated, low-power processors that share processing tasks across hundreds of processors rather than channeling the workload through just a few processors. The method of increasing performance is sometimes referred to as "scaling out."Content Continues Below
When properly implemented, an ARM server can deliver greater processing power while using less energy and requiring less cooling than similar technology based on x86-class processors, which are important considerations for most enterprise data centers. Current ARM server development efforts include Hewlett-Packard's Project Moonshot server, which is based on an array of Calxeda’s EnergyCore ARM Cortex processors, and HP's Redstone servers, which are based on arrays of Intel Atom processors.
ARM servers are not yet direct replacements for conventional x86-class servers. The most important difference is that ARM processors (and the servers that use them) only support limited instruction sets. This limitation means that operating systems and applications must be written specifically for the ARM architecture (such as Android or iOS), and optimized to utilize the ARM processors in a cooperative manner. Microsoft has announced that Windows 8 will include support for ARM chips, which may extend to support for ARM-based servers as well as purpose-built Windows 8 devices. ARM servers will likely fill more niche roles in the data center, such as running web servers or social media applications.