Deployment Requirements
This page lists the hardware and software requirements for deploying a Synnax cluster to production.
Operating System
A Synnax cluster can run on Linux, macOS, Windows, or a containerized environment.
For most production scenarios, we recommend running a glibc-based Linux distribution and a Linux kernel version from the last 5 years, such as Ubuntu, Red Hat Enterprise Linux, or CentOS.
Device Driver Compatibility
If you plan on using our device drivers to integrate directly with your hardware, only certain operating systems are supported.
- National Instruments - Windows, Linux
- LabJack - Windows, Linux, macOS
- OPC UA - Windows, Linux, macOS, or a containerized environment
Hardware
Cluster hardware requirements vary depending on your data acquisition and control workload.
Calculating Workload
The first step in sizing a system to meet your requirements is to calculate the total sample throughput of your workload.
- The approximate number of channels you’ll be acquiring data from. This includes sensors, actuators, logs, and any other channels you may be using.
- The average rate at which you’ll be writing data for each channel. To be even more precise, separate your channels into groups with specific sampling rates.
The total throughput is the number of channels multiplied by the average rate at which you’ll be writing data for each channel.
For example, if we have three groups of channels:
- 10 high speed at 25 kHz -> 250K samples/s
- 100 medium speed at 100 Hz -> 10K samples/s
- 20 actuators sending states at 50 Hz -> 1K samples/s
The total throughput for our cluster is 250K + 10K + 1K = 261K samples/s.
Workload-Based Minimum Requirements
Once you’ve calculated your total throughput, use the following table to determine the minimum hardware requirements for your cluster.
Throughput (samples/s) | CPU (Cores) | Memory (GB) | Storage (GB) | Storage Throughput (MB/s) |
---|---|---|---|---|
250K | 4 | 8 | 100 | 120 |
500K | 6 | 16 | 200 | 1000 |
1M | 12 | 16 | 500 | 600 |
2M | 16 | 32 | 1000 | 1200 |
4M | 24 | 64 | 2000 | 2400 |
When operating Synnax in mission critical environments, we recommend provisioning hardware that is at least twice the minimum requirements. This leaves ample room for workload growth and unexpected spikes in throughput.
Storage
We highly recommend using a high-performance M.2, NVMe, or SATA SSD for storage. This recommendation becomes more of a requirement if you have any very high throughput data acquisition channels (10 kHz or higher).
CPU
The Synnax cluster is CPU intensive during data ingestion, streaming, and control operations. Modern multi-core processors perform best, and are a must when working in safety/performance critical environments.
For production deployments, we recommend using a minimum four core processors from one of the following families:
- Intel x86-64 processors (Xeon or Core i7/i9 series).
- AMD x86-64 processors (EPYC or Ryzen series).
- Apple Silicon Chips on MacOS (M1/M2/M3 series).
Avoid single or dual core processors for anything beyond development or testing environments.
If you plan on using large numbers of calculated channels, increase the number of cores for your deployment.
Memory
The Synnax cluster is generally less memory intensive than it is storage or compute intensive. For most production deployments, we recommend using a minimum of 8 GB of RAM.
Network
Synnax nodes communicate using TCP and/or UDP protocols for cluster coordination, data replication, and client connections. Given Synnax’s high-performance capabilities for both data acquisition and control operations, the network infrastructure must be designed to handle substantial throughput with minimal latency.
Network Interface Requirements
For production deployments, we recommend the following network interface specifications:
Minimum Requirements
- 10 Gigabit Ethernet (10GbE) network interface cards (NICs)
- Full-duplex operation support
- Hardware-based packet processing capabilities
High-Performance Scenarios
- 25 Gigabit Ethernet (25GbE) or higher for workloads exceeding 2M samples/s
- Multiple NICs for redundancy and load balancing in mission-critical environments
- RDMA-capable NICs for ultra-low latency applications
Switch Infrastructure
Network switches should be selected and configured to minimize latency and maximize throughput:
- Low-Latency Switches -> Use enterprise-grade switches with sub-microsecond port-to-port latency.
- Non-Blocking Architecture -> Ensure switches provide full line-rate forwarding across all ports.
- Buffer Management -> Configure appropriate buffer sizes to handle bursty traffic patterns common in data acquisition workloads.
- Quality of Service (QoS) -> Implement traffic prioritization to ensure control commands receive priority over bulk data transfers.
Cabling and Physical Infrastructure
For optimal performance, especially over longer distances:
- Fiber Optic Cabling -> Use single-mode or multi-mode fiber for runs longer than 100 meters.
- Direct Attach Cables (DAC) -> Acceptable for short runs (< 10 meters) between adjacent equipment.
Network Security and Isolation
Critical Security & Performance Requirement -> Never use public internet connections for control operations. Control systems must operate on isolated, dedicated networks.
For production control environments:
- Isolated Networks: Deploy Synnax clusters on dedicated, air-gapped networks separate from corporate IT infrastructure.
- Network Segmentation: Use VLANs or physical separation to isolate control traffic from data acquisition traffic.
- Firewall Protection: Implement network firewalls with strict access control policies.
- VPN Access: When remote access is required, use industrial-grade VPN solutions with multi-factor authentication.