Asynchronous Agents

An asynchronous agent (or just agent) is an application component that works asynchronously with other agents to solve larger computing tasks. Think of an agent as a task that has a set life cycle. For example, one agent might read data from an input/output device (such as the keyboard, a file on disk, or a network connection) and another agent might perform action on that data as it becomes available. The first agent uses message passing to inform the second agent that more data is available. The Concurrency Runtime task scheduler provides an efficient mechanism to enable agents to block and yield cooperatively without requiring less efficient preemption.

The Agents Library defines the concurrency::agent class to represent an asynchronous agent. agent is an abstract class that declares the virtual method concurrency::agent::run. The run method executes the task that is performed by the agent. Because run is abstract, you must implement this method in every class that you derive from agent.

Agent Life Cycle

Agents have a set life cycle. The concurrency::agent_status enumeration defines the various states of an agent. The following illustration is a state diagram that shows how agents progress from one state to another. In this illustration, solid lines represent methods that you call from your application; dotted lines represent methods that are called from the runtime.

Agent State Diagram

The following table describes each state in the agent_status enumeration.

Agent State

Description

agent_created

The agent has not been scheduled for execution.

agent_runnable

The runtime is scheduling the agent for execution.

agent_started

The agent has started and is running.

agent_done

The agent finished.

agent_canceled

The agent was canceled before it entered the started state.

agent_created is the initial state of an agent, agent_runnable and agent_started are the active states, and agent_done and agent_canceled are the terminal states.

Use the concurrency::agent::status method to retrieve the current state of an agent object. Although the status method is concurrency-safe, the state of the agent can change by the time the status method returns. For example, an agent could be in the agent_started state when you call the status method, but moved to the agent_done state just after the status method returns.

Methods and Features

The following table shows some of the important methods that belong to the agent class. For more information about all of the agent class methods, see agent Class.

Method

Description

start

Schedules the agent object for execution and sets it to the agent_runnable state.

run

Executes the task that is to be performed by the agent object.

done

Moves an agent to the agent_done state.

cancel

If the agent was not started, this method cancels execution of the agent and sets it to the agent_canceled state.

status

Retrieves the current state of the agent object.

wait

Waits for the agent object to enter the agent_done or agent_canceled state.

wait_for_all

Waits for all provided agent objects to enter the agent_done or agent_canceled state.

wait_for_one

Waits for at least one of the provided agent objects to enter the agent_done or agent_canceled state.

After you create an agent object, call the concurrency::agent::start method to schedule it for execution. The runtime calls the run method after it schedules the agent and sets it to the agent_runnable state.

The runtime does not manage exceptions that are thrown by asynchronous agents. For more information about exception handling and agents, see Exception Handling in the Concurrency Runtime.

Example

For an example that shows how to create a basic agent-based application, see Walkthrough: Creating an Agent-Based Application.

See Also

Concepts

Asynchronous Agents Library