D5.3 Executive Summary

Mas2tering is developing a multi-agent system (MAS) based ICT solution that enables flexibility management within the low-voltage part of the electricity distribution network where decentralized decision making will bring value and competitiveness. In this vision, communications and in particular agent-to-agent communication plays a key role. In the core of the Mas2tering solution, there is a distributed multi-agent platform in which agents, representing different stakeholders, deployed in different locations, need to communicate with each other. Also, special attention needs to be paid to the connection of this multi-agent platform with non-agent components such as forecasting and monitoring services and Home Area Network (HAN) devices. This report is a design document that addresses all these communication aspects, providing the design specifications required for the implementation of the identified communication components in D5.4 and D5.5. In more detail, key results from this report are highlighted as follows:

  • The design of a multi-agent communication architecture: a two-level architecture is formulated in which devices at home level are connected using ZigBee to create a local network, whereas the other agents are connected over the internet. There are four agent types in Mas2tering, namely: Device, Customer Energy Management System (CEMS, Aggregator (AGR) and Distributor System Operator (DSO). Agent type locations and the possible communication paths summarised in the following table:


  • The definition of a FIPA-ACL compliant communication specification, which has been implemented following a semi-automated model transformation workflow form a Web Ontology Language (OWL) ontology to a dedicated Mas2tering JADE ontology.
  • The communication design to integrate Mas2tering (non-agent based) supporting components (initially assessed in D5.1):
    • Security components
      • With PANDA designed by the Telecommunications Software and System Group (TSSG) for data access control in the Customer premises
      • With the Secure Smart Control component designed by Airbus – formerly Cassidian Cyber Security (CCS) – that adds encryption and authentication capabilities preventing any corrupted agent from communicating within the MAS platform
      • With the Anonymization service designed by Cardiff University (CU) that supplies de-identification services, while accessing raw data in a trusted role from the data historical server.
    • Monitoring components
      • With the Cloud synchronisation OSGi bundle designed by Telecom Italia (TI) that controls the reception and storage of data from the devices as well as obtaining the data from the cloud
    • Forecasting components
      • With forecasting services designed by Cardiff University (CU) that forecast the consumption/production of the devices via web service interface.
  • The design of eleven agent interaction protocols that cover the agent-to-agent communication for the three Mas2tering use cases. For each agent communication protocol, the following aspects are identified:(1) the agent type initiator of the message exchange; (2) the agent type receiver of this initial message exchange; (3) the protocol purpose/parameters; (4) all necessary ACL messages along with the definition of the possible ordering of exchange (i.e. sequence diagram) and (5) security/privacy aspects to be
  • A traceability matrix that maps each of the identified communication requirements with
    the communication component that satisfies it.

In summary, this report provides a detailed design and specifications of the Mas2tering multi-agent communication architecture. Next steps in the Mas2tering communication development work will focus on the completion of the implementation phase of the identified communication components and in particular of: (1) the Device agent that acts as interface between the Energy Box and the rest of the multi-agent platform; (2) the agent-to-agent communication protocols; and (3) the agent communication behaviors that will produce/process the content of the ACL messages.

D5.3 Table of Contents

  • Executive summary
  • Document Information
  • Table of Contents
  • List of figures
  • List of tables
  • Abbreviations
  • Definitions
  • 1 Introduction
  • 1.1 Relationship to other deliverables
  • 1.2 Document Structure
  • 2 Technology Background and Positioning in Mas2tering platform
  • 2.1 Summary of relevant communication protocol standards and the related impact on MAS communication design
  • 2.2 Summary of relevant smart grid information models standards and the related impact on MAS communication design
  • 2.3 Communication requirements
  • 3 Communication Architecture
  • 3.1 Global Communication Architecture
  • 3.1.1 Logical view
  • 3.1.2 Physical view
  • 3.2 Integration with JADE communication
  • 3.2.1 Agent unique identifiers
  • 3.2.2 Agent distributed communication middleware
  • 3.2.3 Agent discovery (the yellow pages service)
  • 3.2.4 Agent interaction protocols
  • 3.2.5 Agent activity logging
  • 4 Data exchange model
  • 4.1 FIPA-ACL compliance
  • 4.4 JADE bean generator
  • 5 Integration of supporting (non-agent) communication components
  • 5.1 PANDA: Access Control to Customer Premises Equipment
  • 5.2 IdM KeyRock GE
  • 5.3 IoT Data Edge Consolidation GE – Cepheus
  • 5.4 Cloud Synchronization Bundle
  • 5.5 Secure Stateless Communication Protocol
  • 5.6 Forecasting services
  • 5.7 Anonymization Service
  • Figure 15 Anonymisation service provision.
  • Figure 16 Anonymisation communication ontology – for the agent-based deployment.
  • Figure 17 Anonymization scenario – forecasting entity submitting a data query
  • 6 Use–cases focused communication design
  • 6.1 Use case 1
  • 6.1.1 Use case description from a communication perspective
  • 6.1.2 Targeted agent interaction protocols
  • 6.1.3 Messages in relation to the data model
  • 6.2 Use case 2
  • 6.2.1 Use case description from a communication perspective
  • 6.2.2 Targeted agent interaction protocols
  • 6.2.3 Messages in relation to the data model
  • 6.3 Use case 3
  • 6.3.1 Use case description from a communication perspective
  • 6.3.2 Targeted agent interaction protocols
  • 6.3.3 Messages in relation to the data model
  • 7 Focus on enhancing reliability and scalability of smartgrid MAS communication
  • 7.1 How Mas2tering communication ensures security and reliability
  • 7.2 How Mas2tering addresses communication requirements
  • 7.3 Extensions and improvements to standard communication protocols and information models
  • 8 Summary and Conclusion
  • References
  • Annex A Ontology concept descriptions

D5.3 Highlights

The Secure Stateless Communication Protocol is a protocol implemented in a solution developed by Airbus (formerly Cassidian Cybersecurity) in order to easily secure stateless communication, such as REST API Communication. This solution does not need any PKI deployment, which is very complex and Expensive to deploy, and manage. this solution is composed of a Secure Communication Module, which secures the communication and a device manager, which registers the allowed devices and allow management through web REST API and Graphical User Interface (GUI). in Mas2tering context, the Secure Communication Module will be embedded in a JADE add-on (JADE-SSCP). A Device Manager module could be embedded in the Energy Box in order to manage agent security configuration. Below is Mas2tering SSCP deployment diagram:


This solution implements two processes:

  • Enrolment phase: in order to be allowed to communicate, the agent needs to send an enrolment request to register itself (Figure 11).


  • Communication phase: once the enrolment phase is finished, the agent can communicate with other agent in a secure way, with authentication and encryption (Figure 12).


D5.3 Conclusion

This document presents the design of the Multi-Agent global communication architecture for the Mas2tering project, providing the design specifications required for the identified communication components. This design process follows the adopted Gaia agent development methodology for Mas2tering, and further advances the Gaia interaction model defined in D2.2 (i.e. with the corresponding high-level interaction scheme) to define the agent communication protocols and ACL messages.

Based on this document, Mas2tering developers can now rely on:

  • A global design of the two-layered multi-agent communication architecture, including a logical view (i.e. specifying the possible paths of the agent communication) and a physical view (i.e. specifying the agents that will be deployed in Mas2tering, their location and location
    of other services).
  • The design of the Mas2tering data exchange model and a methodology to generate the messages exchanged from the Mas2tering ontology. The proposed model is aligned with relevant standards (in particular with FIPA-ACL, energy@home, CIM, IEC 61968 and OpenADR). A methodology to create automatic JADE messages is proposed including a mapping from standard vocabularies to JADE compatible source code and the definition of a JADE bean generator for automatic code generation.
  • The communication design of the Mas2tering MAS platform integrating other non-agent communication (security, forecasting, monitoring) components.
  • A mapping of this global architecture to the project use cases with the definition of the communication details in each use case.
  • The specification of eleven agent interaction protocols that cover the agent-to-agent communication for the three Mas2tering use cases.

In parallel, the design of the communication architecture has led to the identification of several development tasks that have been linked in Pacific project management software to the partners in charge of their implementation. These developments tasks are:

  • The implementation of a Device agent that acts as an interface between the Telecom Italia Energy Box (i.e. running JEMMA software) and the rest of the multi-agent platform.
  • The implementation of the generic communication components that define the global communication architecture and of the agent communication protocols (in JADE, based on FIPA-ACL) for each use case.
  • The implementation of agent communication behaviours that will produce/process the content of the ACL messages (following the FIPA ACL Message Structure Specification), thus exploiting the JADE content language and ontology support with connection to the Mas2tering ontology.

The current status of FIWARE Generic Enablers presents challenges to integrating them into the Mas2tering platform due to their continued problems of stability, reliability and maturity. Mas2tering has identified alternative components that provide more reliable and stable reusable functions. Such components include PANDA, ARX anonymization and the JADE framework. However, we include the description of the integration specifications of two Generic Enablers that have the potential of adding useful features to Mas2tering.

The next phase in communication development work will include the completion of the implementation of these communication components which is already underway.

D5.3 Download Link