AppGate ZTNA addresses traditional high availability and load balancing challenges through several key design principles.
Appliance based
All system functions run on Ubuntu-based appliances that you can deploy as cloud instances, virtual machines, or physical appliances. The software is optimized for multi-core processors and built-in hardware features such as AESNI. Gateways can achieve throughput in the range of 20 Gb/s to 40 Gb/s, even on virtual hosts.
HA Controllers
An AppGate ZTNA Controller handles on-demand transactions using REST calls, which prevents a bottleneck for signed-in users who connect to and push traffic through Gateways. As a result, short periods of Controller unavailability at a Site may go unnoticed by users unless their tokens expire during that time.
For high-transaction environments, you can specify a larger Controller to improve performance. There is a practical limit to the number of cores and threads a single appliance can use. When you reach this limit, AppGate ZTNA supports HA Controllers, which enable load-balanced, highly available, multi-master Controllers.
Site-based access
When a Collective contains multiple Sites, AppGate ZTNA distributes Gateway load based on policy. Entitlement tokens are issued on a per-Site basis according to assigned policies. This allows you to provision Sites in Europe for European employees and Sites in the Americas for American employees. Clients do not access Sites that have not been explicitly allowed.
Gateway Micro-firewall service
AppGate ZTNA uses a patented approach that allows the Gateway to provide a secure tunnel service and a micro-firewall instance for each connected client using a separate thread. Once the client establishes a secure tunnel with the Gateway, it pushes the user entitlement token to initiate access. The Gateway uses the entitlement token to set up an individual firewall table for each micro-firewall engine. By distributing micro-firewall services across all processing cores, AppGate ZTNA always assigns new client connections to a processor with available resources. Each firewall service manages only a few rules for the individual user.
HA Gateways
AppGate ZTNA Gateways are designed for scalability. Adding HA Gateways to a Site, or subdividing Sites into smaller micro-sites, scales infrastructure performance linearly. Sites and Gateways are independent, and AppGate ZTNA requires no inter-Gateway traffic. Adding a new Gateway to an existing Site, or adding a Site protected by a new Gateway, does not affect the performance of existing appliances, because the system relies on token passing rather than inter-appliance traffic. There is no operational limit to the number of Gateways you can add to manage performance and reliability as network traffic or user population grows.
Load balancing
Use the internal HA mechanisms for load balancing on both Controllers and Gateways. This ensures client connections are distributed optimally. For example, you can use weightings to ensure that Gateways with 8 vCPUs receive twice the number of connections as those with only 4 vCPUs.