Mobile-C: a mobile agent platform for mobile C/C++ agents

Bo Chen, Harry H. Cheng, Joe PalenSoftware 2006 · DOI

TL;DR

This paper introduces Mobile-C, a FIPA-compliant agent platform designed for C/C++ mobile agents. It addresses the need for flexible, interoperable distributed systems by enabling agents written in C/C++ to migrate between platforms, leveraging an embedded C/C++ interpreter (Ch) and FIPA ACL messages for communication and migration.

What Problem Does This Paper Address?

The paper tackles the challenge of building flexible, interoperable, and dynamic distributed systems, particularly those involving low-level hardware control and real-time data processing. Existing middleware solutions like CORBA and web services, while good for general distributed computing, often fall short for highly dynamic, time-critical, and embedded environments. Agent-based computing offers a promising alternative, but many existing mobile agent systems are either not interoperable (lacking adherence to standards like FIPA) or are primarily Java-based, making direct interfacing with C/C++ systems difficult. The authors aim to bridge this gap by providing a FIPA-compliant platform that supports mobile agents written in C/C++, enabling seamless integration with existing C/C++ codebases and hardware.

Methods

The core of Mobile-C is its FIPA compliance, ensuring interoperability with other FIPA-compliant systems. To support mobile C/C++ agents, the platform integrates an embeddable C/C++ interpreter called "Ch" (from Softintegration, Inc.). This interpreter acts as the Agent Execution Engine (AEE), allowing mobile agents written in C/C++ to run within the platform.

Key methodological contributions include:

  • C/C++ as Mobile Agent Language: This choice facilitates direct interfacing with low-level software modules and hardware, a significant advantage for systems like traffic management.
  • Embeddable C/C++ Interpreter (Ch): Ch provides a runtime environment for C/C++ agents, offering portability and security. It also allows for easy integration with the agent platform's native C/C++ code.
  • Agent Mobility Protocol: Mobile-C defines a protocol for agent migration that leverages FIPA Agent Communication Language (ACL) messages. This means agent communication and migration share the same underlying mechanism.
  • Weak Migration: Due to limitations in capturing the full execution state of a C/C++ process with the chosen interpreter, Mobile-C currently supports weak migration, where only the agent's data state is transferred. Agent tasks are managed via a task list and a progress pointer.
  • XML Encoding: FIPA ACL messages, including those for agent migration, are encoded in XML, promoting flexibility and ease of parsing.

Datasets

Not discussed in the paper. The paper focuses on the platform design and a simulated application.

Key Insights and Ideas

  • Bridging the C/C++ Gap for Mobile Agents: The most significant insight is the successful integration of C/C++ as a mobile agent language within a FIPA-compliant framework. This opens up agent technology to a vast ecosystem of existing C/C++ code and systems.
  • Unified Communication and Migration: Using FIPA ACL messages for both agent communication and migration is a clever design choice. It simplifies the platform architecture and reduces development effort by leveraging a single, standardized communication mechanism.
  • Small Footprint and Embeddability: The choice of Ch as the interpreter and the overall design philosophy of Mobile-C emphasize a small footprint, making it suitable for resource-constrained environments where traditional Java-based agent platforms might be too heavy.
  • Agent-Based Traffic Management Potential: The application to highway traffic detection and management demonstrates the practical utility of mobile agents for dynamic code deployment, real-time data fusion, and performing unanticipated actions in complex, distributed systems.

Key Results

The paper presents the design and implementation of the Mobile-C platform and demonstrates its application in a simulated highway traffic detection and management system (ABRTTDMS).

  • FIPA Compliance: Mobile-C is designed to be compliant with IEEE FIPA standards, ensuring interoperability.
  • Mobile Agent Migration: The platform successfully implements agent migration using FIPA ACL messages encoded in XML.
  • C/C++ Agent Execution: Mobile agents written in C/C++ can be executed and migrated across different platforms.
  • Application in ITS: The ABRTTDMS simulation shows that mobile agents can be used for dynamic code deployment and performing unanticipated actions, enhancing flexibility and adaptability in traffic management. The system demonstrated the ability to query distributed databases for real-time traffic information.

Limitations and Future Work

  • Weak Migration: The current implementation of Mobile-C supports only weak migration due to limitations in capturing the full execution state of C/C++ processes with the embedded interpreter. The authors suggest that future work could explore methods to achieve strong migration.
  • Execution Speed: While C/C++ offers advantages in interfacing with hardware, interpreted execution can lead to slower performance compared to compiled native code. The authors acknowledge this but argue that for systems where major computation is handled by stationary agents, the impact of mobile agent interpretation is mitigated.
  • XML Processing: The paper mentions that mobile agents can directly process XML data using an interpretive software package (Ch XML), which is presented as an advantage for forming programmable information bases. This feature is noted as being presented in a separate paper.
  • Future Work: The authors suggest that Mobile-C has potential for applications in agile manufacturing, autonomous sensor fusion, and intelligent transportation systems. Further development could explore these areas and potentially enhance the migration capabilities.

Summary updated

Abstract

This article presents the design, implementation and application of Mobile-C, an IEEE Foundation for Intelligent Physical Agents (FIPA) compliant agent platform for mobile C/C++ agents. IEEE FIPA standards are a set of specifications designed to ensure the interoperation between agents in a heterogeneous network. Mobile-C conforms to the FIPA standards both at agent and platform level. Mobile-C extends FIPA standards to support mobile agents by integrating an embeddable C/C++ interpreter—Ch—into the platform as a mobile agent execution engine and defining a mobile agent mobility protocol to direct agent migration process. Agent migration in Mobile-C is achieved through FIPA agent communication language (ACL) messages encoded in XML. Using FIPA ACL messages for agent migration provides a straightforward but effective way for inter-platform agent migration in FIPA compliant agent systems as both agent communication and migration can share the same communication mechanism. Choosing scriptable C/C++ as a mobile agent language allows mobile agents easy interfacing with low-level software modules and underlying hardware. Mobile-C has been used to simulate highway traffic detection and management. The agent-based traffic detection and management system uses agent technology for real-time distributed traffic information fusion. Mobile agents in the system are used for dynamic code deployment and performing unanticipated actions. The application of agent technology shows a great potential for enhancing the interoperability, flexibility and distributed computing capabilities of intelligent transportation systems. Copyright © 2006 John Wiley & Sons, Ltd.
BibTeX
@inproceedings{chen2006mobile,
  author = {Bo Chen and Harry H. Cheng and Joe Palen},
  title = {{Mobile-C: a mobile agent platform for mobile C/C++ agents}},
  booktitle = {Software},
  year = {2006},
  doi = {10.1002/spe.742},
}
Figure