What are the system requirements of the virtual machine or computing device on which the gateway software is installed?

To determine the system requirements for the gateway, the first step is to estimate how many devices will be sending traffic to it and which high-availability model you will use. See, How will load balancing work if a resource can be reached by multiple gateways?

Automatic high-availability will allow for horizontal scaling allowing you to add gateways to meet the needs of the traffic.

To determine the system requirements of the gateway’s computing instance, follow the steps below:

1. Determine how many devices this gateway should be capable of handling simultaneously. Let us refer to this number as Nt.

2. Estimate the average bandwidth of traffic from a device based on the private/SaaS/internet resources the particular gateway provides access to. This can be based on network traffic observation when the device usage can be monitored or based on studies, such as this one which provides bandwidth recommendations by different application categories. Let us refer to this number as Ub.

3. The network interface bandwidth that the gateway instance should support (Nb) = 1.2 * Nt * Ub.

4. Choose a VM type that has the capacity to support Nb Mbps of traffic. Note that some IaaS and VPS providers artificially limit data rate of the network interface.