Scale Computing was founded in 2009, building a scale-out, software-defined storage system that’s used as the foundation for their HC3 Hyperconverged Infrastructure (HCI) solution. The HC3 platform is a cluster of three or more single- node systems that run the company’s “HyperCore” software at the Linux kernel level, not as a VM like most other HCI solutions. Each node runs a built-in hypervisor, a KVM-derivative, that’s included with the system, along with a migration utility to move VMs from other hypervisors. In fact, there is no choice to use VMware, Citrix or Hyper-V.
An HC3 cluster is comprised of three to sixteen 1U or 2Unodes, each containing up to 12 hard disk drives, 4 SAS SSDs or 10 NVMe SSDs. Storage configurations include hard drive only, hybrid and all-flash. Scale Computing’s solution supports flash storage as a tier, not a cache, and features an automated tiering software that moves data based on priorities set for each VM.Scale Computing also has nodes with GPUs. The HC1250DFG supports up to two and the HC5250DFG, up to five NVIDIA Tesla T4 GPUs.
The HE 500 series of nodes are single-CPU, 4-drive nodes designed for smaller locations or remote offices. These lower-cost models are available in 1U rack or tower form factors. The HE150 runs the complete Scale Computing software stack on the Intel NUC, a 4” x 4” with a single SSD and up to 6 CPU cores. This is currently the smallest node available from a major HCI vendor.
HyperCore is designed around the company’s proprietary data distribution architecture “SCRIBE” (Scale Computing Reliable Block Engine) that spreads data blocks of selected redundancy around the cluster in a variable RAID 10 type of protection. SCRIBE presents these blocks directly up to the hypervisor, eliminating additional file system layers in the stack to maximize performance and simplify data management. This architecture also reduces total resource requirements which enables HC3 to run with lower core-count CPUs and less memory.
Download the full Product Brief now!