Agent pools are logical collections of Sipfront agents. When you configure a test, you select the pool in which to run the test, and the Sipfront orchestration system will auto-select agents that are registered in that pool.
Public vs private pools¶
There are several public agent pools that we as Sipfront provide as part of the service, and you can also create our own private pools and start your own local agents, e.g. on your own laptop or on a VM or bare metal machine in your premises.
At Sipfront, we provide public agents which you can use as part of the service, and we group them in agent pools which are named according to their geographic location (e.g. aws-eu-central-1 or aws-us-east-1). In general, there is no hard requirement how agents in these pools are grouped.
You can very well name your agent pool my local test pool, and in that agent pool start one agent on a machine in Australia, the other one somewhere in the US and the third one on your local laptop.
Agent pools are a logic collection of agents. Agents in a pool can technically be located anywhere in the world.
Sipfront agent pools follow geographic naming conventions, so you will know up-front in which region they are running.
Anatomy of an Agent¶
Sipfront agents are Docker containers either run on Sipfront's cloud infrastructure or on your own premises, grouped into Agent pools.
Internally, they run an agent process, which communicates with the Sipfront orchestration system to receive commands to execute, and to send metrics for analyzing test states.
Within the agent, we utilize various tools such as kamailio, rtpengine, baresip and sipp to generate and terminate - based on the test scenario - the test traffic.