Skip to main content
cubecos
Version: 3.0

Deployment Considerations

This section outlines the key hardware and network requirements for deploying CubeCOS across various cluster configurations. Each component is critical in ensuring a stable, performant, and scalable platform, from CPU and memory planning to storage media selection and network architecture.

Processors (CPU)​

Processor selection for CubeCOS nodes should align with each node's computing role and workload demands. Refer to the hardware specifications section for minimum requirements, and size compute nodes based on the specific performance needs of the intended workloads.

Memory​

The minimum memory ensures stable operation of the CubeCOS infrastructure, while additional memory may be needed based on workload demands.

SDS reservation​

Plan for memory reservations, including 4 GB per storage device for optimal SDS (Ceph OSD) performance.

warning

Intel Optane Memory modules are not supported.

Storage​

Boot media​

For redundant and reliable boot disk configurations, use the following:

  • Dell BOSS cards (RAID1)
  • HPE NVMe Hot-plug boot solutions (RAID1)
  • Any other vendor provided redundant (RAID1) and boot-optimized storage media.

These options ensure fail-safe boot operations in both production and edge deployments.

Storage media​

CubeCOS uses Ceph, a Software-Defined Storage (SDS) solution, to provide a scalable, distributed, and highly available storage backend.

Storage controller​

Storage controller type and configuration play a critical role in cluster performance. The following are recommended configurations:

  • Use Host Bus Adapters (HBAs) or RAID controllers in IT mode for reliable, streamlined deployment in distributed storage setups.
  • Avoid mixed-mode controllers that offer both RAID and passthrough modes, as they may cause instability or degraded performance.
  • For NVMe storage, CubeCOS supports direct NVMe access without any controllers.
tip

Avoid mixed-mode storage controllers that support RAID and passthrough, as they may introduce instability.

Supported storage media​

  • NVMe (U.2, M.2)
  • NL-SAS
  • SAS SSD/HDD
  • SATA SSD/HDD

Hybrid Storage Deployment​

Hybrid storage configurations combine fast flash media for caching read operations with HDDs for bulk storage capacity.

To maintain optimal performance and prevent storage bottlenecks, use a 1:10 raw cache-to-storage ratio when deploying hybrid storage configurations.

Network​

CubeCOS relies on various software-defined services that require stable and high-availability network connections to maintain cluster health, data consistency, and service synchronization.

Basic requirements​

To meet these requirements, ensure the following requirements are met:

  • LACP bonding (Protocol: 802.3ad, Rate: Fast, Xmit Policy: Layer 3+4) for link aggregation.
  • Highly available switching architectures to ensure failover resilience and performance.
  • Higher interface speeds in environments with intensive storage throughput.
Info 💡

As faster storage media (e.g., NVMe) are used, matching network performance becomes critical to support replication and synchronization workloads.

Traffic steering and separation​

Adaptable network segmentation and flow control are designed to support diverse deployment architectures and meet varying performance requirements.

  • A dual-port 10Gb NIC can handle both user and internal cluster traffic for low-throughput environments.
  • For higher-performance or production deployments, Bigstack recommends traffic separation across:
    • Management
    • User access
    • Data
    • Storage networks

Traffic separation can be implemented in two ways:

  • VLAN-based trunking, where tagged traffic is separated over a single physical uplink.
  • Dedicated physical interfaces may be required for compliance or guaranteed performance.
info

Network configuration can be aligned with existing client infrastructure and policies for seamless integration.

NVMe and High-Speed Networking​

When deploying NVMe storage, network performance is a critical factor.

  • Minimum recommended speed: 25GbE
  • Preferred speed for optimal performance: 100GbE

This ensures the network can sustain the IOPS and bandwidth demands of high-speed storage replication and service orchestration.