Video appliance requirements #
Use these requirements when preparing a Linux machine to run a video appliance with the CLI tool. The machine can be a physical server or a virtual machine. For control-plane installation, follow the requirements in your deployment guide.
Operating system and Docker #
Use an AMD64 (x86-64) Linux host with Docker Engine. Docker Desktop is not supported for this deployment. Docker-hosted appliances still need Docker even when the Nimbra Edge control plane runs on Kubernetes.
Install Docker Engine 25 or later using the Docker installation instructions. The CLI tool configures the appliance containers and host settings; you do not need to assemble the containers yourself.
Time synchronization #
Enable NTP or another time synchronization service on every appliance. Accurate, synchronized clocks are needed for reliable video transport and monitoring.
Firewall support #
The Linux kernel must support nftables. Video appliances and Edge Connect use it to restrict traffic to the source address ranges configured for inputs and outputs. On distributions where the modules are not loaded automatically, ensure that nf_tables and the required protocol modules are available.
Connectivity #
The appliance must be able to reach the Nimbra Edge API over HTTPS (TCP 443), resolve its hostname, and reach its time server. Allow the configured video tunnel ports between appliances; the default range is UDP 20000–21000. Also allow the ports used by your inputs and outputs.
For deployments behind NAT, configure the region’s address mappings so other appliances can reach the video hosts. Your firewall must allow traffic to those public addresses and forward it to the corresponding video hosts.
Hardware requirements #
Start with the following resources for each video appliance. Actual capacity depends on the number of streams, their bitrates, and any processing you enable.
| Resource | Requirement |
|---|---|
| CPU | 4 cores |
| Memory | 8 GB |
| Storage | 20 GB |
Give video workloads predictable CPU and network capacity. On a virtualized host, avoid competing workloads that can interrupt video processing.