Compute servers

Use this section to get a detailed understanding and configuration details of compute servers that are used in the IBM Fusion HCI appliance.

Compute-only nodes are designed for applications that require high processing power and can use external storage solutions. These nodes are ideal for workloads that depend on shared storage or have low storage requirements and provide the necessary resources for workloads to run efficiently. Also, the compute servers are used a control nodes and worker nodes in the OpenShift® cluster.

The available compute servers are:
  • Dell hardware - 9155-D10 and 9155-D14. For configuration details, see Dell hardware.
  • Lenovo hardware - 9155-C10 and 9155-C14. For configuration details, see Lenovo hardware.
You can have either three 9155-C10/D10 and 9155-C14/D14 servers with storage as a minimum configuration, and also you can expand your IBM Fusion HCI beyond the minimum three servers by adding combinations of models C10/D10 and D10/D14. Each rack can support up to 16 servers. Each compute server can have a minimum of two storage drives that can be increased up to a maximum of ten storage drives.
Important: You cannot mix and match different drive capacities within the same rack. Choose one capacity and use it consistently throughout.

Key capabilities

The following are the key capabilities of the compute server:
  • Compute servers offer customizable CPU and memory allocation, allowing you to tailor resources to meet specific workload requirements.
  • You can adjust the amount of CPU and memory in each server to optimize performance and cost.
  • Compute servers can be configured as storage-rich servers, providing additional storage capacity for workloads that require it.

Advantages

The following are the key benefits of the compute servers:
Improved resource utilization
By customizing CPU and memory allocation, you can ensure that workloads have the necessary resources to run efficiently, reducing waste and optimizing utilization.
Enhanced flexibility
Compute servers can be scaled up or down to accommodate changing workload demands, making it easier to adapt to evolving business needs.
Cost effective
With the ability to customize resources, you can optimize costs by allocating only the necessary resources for each workload, reducing unnecessary expenses.

Hardware configuration

Dell
9155-D10 (32 core)
  • Dell R660XL
  • 2x 16-core Intel Xeon 4th Gen Gold 6426Y 2.5GHz 185W processors
  • 256GB RAM (16x 16GB DDR5 DIMMs)
  • Optional memory upgrade to 512GB (32x 16GB DIMMs)
  • 0-10x 7.68TB or 3.84TB PCIe NVMe drives
  • 1x NVIDIA ConnectX-6 DX dual-port 100GbE PCIe NIC
  • 1x Intel I350 1GbE RJ45 4-port OCP Ethernet Adapter
  • 2x 960GB M.2 NVMe OS drives (RAID 1)
9155-D14 (64 core)
  • Dell R660XL
  • 2x 32-core Intel Xeon 4th Gen Gold 6438N 2.0GHz 205W processors
  • 1024GB RAM (16x 64GB DDR5 DIMMs)
  • Optional memory upgrade to 2048 GB (32x 64GB DIMMs)
  • 0-10x 7.68TB or 3.84TB PCIe Gen 5 NVMe drives
  • 1x NVIDIA ConnectX-6 DX dual-port 100GbE PCIe NIC
  • 1x Intel I350 1GbE RJ45 4-port OCP Ethernet Adapter
  • 2x 960GB M.2 NVMe OS drives (RAID 1)
Lenovo
9155-C10 (32 core)
  • Lenovo SR630 V3
  • 2x 16-core Intel Gold 6426Y 2.5GHz 185W processors (“Sapphire Rapids”)
  • 256GB RAM (16x 16GB DDR5 DIMMs), expandable to 512GB RAM
  • 0-10x 7.68TB or 3.84TB PCIe generation 5 NVMe drives
  • 1x NVIDIA ConnectX-6 DX dual-port 100GbE PCIe NIC
  • 1x NVIDIA ConnectX-6 LX dual-port 25GbE PCIe NIC
  • 1x Intel I350 1GbE RJ45 4-port OCP Ethernet Adapter
  • 2x 960GB M.2 NVMe OS drives (RAID 1)
9155-C14 (64 core)
  • Lenovo SR630 V3
  • 2x 32-core Intel Gold 6438 2.0GHz 205W processors (“Sapphire Rapids”)
  • 1024GB RAM (16x 64GB DDR5 DIMMs), expandable to 2048GB RAM
  • 0-10x 7.68TB or 3.84TB PCIe generation 5 NVMe drives
  • 1x NVIDIA ConnectX-6 DX dual-port 100GbE PCIe NIC
  • 1x NVIDIA ConnectX-6 LX dual-port 25GbE PCIe NIC
  • 1x Intel I350 1GbE RJ45 4-port OCP Ethernet Adapter
  • 2x 960GB M.2 NVMe OS drives (RAID 1)