Raspberry Pi Cluster

I have worked most of my professional life with systems built with Intel x86, x64 and ia64 chips. Intel processors use Complex Instruction Set Computing while Arm uses Reduced Instruction Set Computing. Processors with CISC architecture look to move complexity to hardware to simlify code, so small code sizes with high cycles per second. Processors with RISC architecture invert the relationship larger code size with low cycles per second.

Anyone who has worked in a data center has an awareness of importance of the amount of heat generated by Intel powered servers. There isn’t much heat generated from allowing electricity to flow through something (unless there is A LOT of electricity). All of Arm’s designs are energy efficient which is why they have become popular for running in smartphones, tablets and other embedded devices. If Arm processors can get traction used in servers this can only be good news for data center electricity usage.

I enjoy learning by doing and to begin to better understand Arm looked to build a cluster of Raspberry Pi. My goals of cluster include:

• Understanding more about Raspberry Pi hardware
• Understanding how an distributed application might work across physical hosts using MPI. MPI is a programming model that is widely used for parallel programming in a cluster.
• Understanding how micro-servies application may work in Edge locations running Kubernetes on low-cost hardware.
• Understanding how ESXi on Arm development is progressing

Project repository: https://github.com/darrylcauldwell/piCluster