Scaling Trusted Agent for Fault Tolerance

More than one Trusted Agent Service can be provisioned within a single private network to allow for increased redundancy of communication with private resources. If one Trusted Agent Service goes down or is otherwise unavailable, additional Trusted Agent Services can facilitate communications with the same private resources.

The Trusted Agent Hub will distribute work among registered Trusted Agents based on a simple selection algorithm that considers the last ping time for each Trusted Agent. As a result, work will not necessarily be load balanced between the Trusted Agents but will be distributed between them depending on the timing of requests and the timing of pings received from each agent.

To enable scale-out of Trusted Agent on a single private network, install and configure more than one Trusted Agent Service on that private network, and then configure each to use the same Trusted Agent Hub URL and Shared Key. No further configuration is required.

While there is no enforced limit on the number of Trusted Agent Servers that can be provisioned and connected to a single Hub, you might experience performance and maintenance issues with large numbers of Trusted Agent Servers. Therefore, no more than 20 Trusted Agent Servers are recommended. Since each Trusted Agent Server hosts five Trusted Agents (one for each feature that uses Trusted Agent), this recommendation allows for 100 Trusted Agents.