Fork me on GitHub

Extending the Janus SRE with Network Connection

Janus is an open-source multi-agent platform fully implemented in SARL. Janus could be used as an agent-oriented platform, an organizational platform, and/or an holonic platform. It also natively manages the concept of recursive agents and holons. Janus platform was initially published during the 2007-2008 period as a pure Java framework. Since 2014, Janus is fully reimplemented to support the SARL Agent-Oriented Programming Language. And since 2020, it is fully re-implemented using the SARL language.

Janus implementation follows the standards of software engineering as much as possible. Basically, the key implementation choices of Janus that are under interest in this tutorial are:

  1. Microkernel architecture: Janus implementation is based on a microkernel that provides the key features of an agent framework, and that could be easily extended. The key features are implemented as services.
  2. Service-based architecture: The main features of the Janus framework are implemented as services, e.g. context management, agent lifecycle, logging, etc.
  3. Dependency Injection: Modules (and services) are injected into the using objects with Google Guice. It provides support for dependency injection using annotations to configure Java objects. Dependency injection is a design pattern whose core principle is to separate behavior from dependency resolution.

By default, Janus does not provide a support for exchanging events over a computer network. For enabling this communication feature, a specific Janus extension is needed.

This documentation page describes the extension io.janusproject.kernel.network that adds the communication among Janus nodes over a computer network.

The extension described on this page was defined according to the standard method for extending Janus.

1. General Principles of the Extension

Basically, the networking feature of the Janus SRE enables you to launch instances of the Janus kernel on different computers, and exchange events between these instances.

General Networking Architecture

In order to implement this feature, the Hazelcast library is used. Hazelcast is an open source in-memory data grid based on Java. In a Hazelcast grid, data is evenly distributed among the nodes of a computer cluster, allowing for horizontal scaling of processing and available storage. Backups are also distributed among nodes to protect against failure of any single node.

In addition to the data-sharing and efficient communication means, Hazelcast provides the ability to build the node network automatically. Each Hazelcast node that is associated to a given cluster name is connected automatically to the other nodes in the same cluster on the local network. It means that you do not need to provide a network configuration to the Janus SRE.

2. Maven Inclusion

In order to include the networking extension into your project, you could add the following Maven dependency.

 <project>
   ...
    <properties>
       ...
       <janus.version>3.0.12.0-SNAPSHOT</janus.version>
    </properties>
    ...
    <dependencies>
       ...
       <dependency>
          <groupId>io.janusproject</groupId>
          <artifactId>io.janusproject.kernel.network</artifactId>
          <version>${janus.version}</version>
       </dependency>
       ...
    </dependencies>
    ...
 </project>

The io.janusproject.kernel.network maven module provides the network extension AND all the libraries related to the Janus SRE.

3. Configuration of the plugin

The following table provides a short description of the configuration parameters that are related to this networking extension. More details could be obtain by calling your Janus-based software with the command-line option -H for example.

Section Property Type Description
srenetwork     Configuration for the SRE networking feature
  clusterName string Name of the Hazelcast cluster. By default, this name is based on the identifier of the agent root context.
  enable boolean Enable or disable the networking features.
  clusterMemberIPs string Specify the list of the IP adresses for the members that are allowed to be in the cluster. This configuration is used only if the join method is configured to TCP/IP.
  joinMethod “multicast” | “tcp_ip” Specify the method for joining the Hazelcast cluster over the computer network
  minClusterSize int Specify the minimal number of nodes to wait for starting in the cluster before really launching the SARL Run-time Environment
  portAutoIncrement boolean Enable or disable the auto-incrementation of the Hazelcast port numbers.

Copyright © 2014-2021 the original authors or authors.

Licensed under the Apache License, Version 2.0; you may not use this file except in compliance with the License. You may obtain a copy of the License.

You are free to reproduce the content of this page on copyleft websites such as Wikipedia.

Generated with the translator io.sarl.maven.docs.generator 0.12.0-SNAPSHOT.