ARM’s Battle for the Datacentre: The Contenders
As enterprise-ready ARM-based servers approach the market, the landscape of data analytics and high-performance computing is set for a major shift.
The first enterprise-ready, ARM-based servers are approaching the market, bringing a new era of "energy-sipping" systems to the datacentre. These 64-bit machines are designed to tackle a far broader range of tasks than the 32-bit experiments of previous years.
Beyond simple web serving, these systems are built to power data analytics on Hadoop clusters, manage NoSQL data stores, handle streaming media, and drive high-performance computing—often sharing processing duties with GPUs, FPGAs, or ASICs.
The Shift to Parallelism
The core advantage of ARM-based servers lies in their ability to process computationally light workloads in parallel across clusters of thousands of cores. These dense clusters of low-power servers handle parallelisable tasks more efficiently than a smaller number of high-powered chips.
The result? Better performance per watt and per square foot of datacentre space—critical metrics for driving down the massive costs of running a large-scale server estate.
The Key Players
While the software ecosystem for ARM web serving and data analytics is rapidly maturing, the hardware is catching up. Major players in the ARM-based server space include:
- Applied Micro: Leading with its X-Gene boards.
- AMD: Branching beyond x86 with its Opteron A1100 processor.
These chips are based on the ARM v8 architecture, which introduces support considered critical for business environments, including 64-bit cores and error-correcting code (ECC) memory.
Looking Ahead
The transition to ARM in the datacentre represents a fundamental shift in how we think about scale. Instead of relying on brute-force power from a few chips, we are moving toward an era of distributed, efficient, and highly parallel processing. At Ekvi Technologies, we monitor these infrastructure shifts closely to ensure our clients' applications are deployed on the most efficient and scalable architectures available.