Architectures -- Projects

Comparison of SW Agent Based Framework w/J2EE Framework

Executable Architecture Methodology for Analysis

Multiagency Enterprise Architecture Planning Framework

Architectures investigates technologies for designing and integrating complex, COTS-based systems of systems that can adapt and evolve with advances in software and information technologies. This area also seeks to define and measure the attributes of a good architecture. Use and evaluation of tools such as Common Object Request Brokers (CORBA) fall into this area.

Comparison of Software Agent Based Framework w/the J2EE Framework

Margaret Lyell, Principal Investigator

Washington

Problem
The problem is to determine interoperabilities, conflicts, and synergies (ICS) between applications in FIPA-compliant software agent frameworks and those developed for the Java 2 Enterprise Edition (J2EE) framework. FIPA, the Foundation for Intelligent Physical Agents, provides standards for software multi-agent systems (MAS). The J2EE is an exemplar of the thin

client, high bandwidth, multi-tier, component-based, Web-centric architecture prevalent today.

Objectives
We ask focused questions to delve into potential ICS. For example: “Can messaging provide for interoperability even though there is a semantic mismatch between applications in the two frameworks?” “How should a software agent system offer or use (from J2EE) a service?” “Can Web Services provide an interoperability bridge?” We seek answers compatible with standards and specifications for both frameworks.

Activities
Completed research threads include: “Messaging as an interoperability bridge,” “Performance study of FIPA-OS,” and “Conceptual comparison of two architectural frameworks: J2EE and FIPA-MAS.” Ongoing threads involve “Web Services as an interoperability bridge,” “Security across J2EE and FIPA-MAS frameworks,” and “Petri net modeling of the combined FIPA-MAS and J2EE system.”

Impacts
We have presented details of our interoperability approaches at FIPA and AgentCities meetings as input into standards activities, and contributed software to the (open source) FIPA-OS community. Presentations and papers have been accepted and given at conferences, including JavaOne, Multi-Agent Interoperability Workshop '02, and Army Science Conference. The project is also collaborating with the High Confidence Software project in a sub-project to build an agent-centric container.

Executable Architecture Methodology for Analysis

Dr. Tom Pawlowski, Principal Investigator

Bedford and Washington

Problem
Current DOD and other federal architecture frameworks describe systems using static representations that limit the user’s ability to analyze system capabilities, particularly the performance and effectiveness of the system in its operational environment. This hinders the utility of the architecture to support key aspects of the systems acquisition process, such as requirements development, analysis of alternatives, systems engineering, and training with the system.

Objectives
The project will develop a methodology to support the analysis of a system using executable architectures. We will integrate an executable operational architecture model of processes and organizations with an executable communications architecture model and a combat simulation to provide a dynamic analytical environment for measuring system performance and effectiveness.

Activities
We will review modeling tools and develop a methodology to dynamically analyze architectures. We will create a prototype tool to implement the methodology by developing the Integrated C4ISR Architecture Management System (ICAMS) to convert architectural specifications into executable representations, and develop a network and communication system architecture. We will establish a federated simulation environment with associated executable specifications, and evaluate the methodology using a practical application.

Impacts
The methodology developed could significantly reduce the time needed to complete an architecture deliverable. It will also provide an easy way of translating operational requirements into technical system requirements, reducing acquisition costs by providing a more cost-effective way of exploring alternative concepts. This project will help steer architecture development efforts toward the use of more rigorous methodologies that support static and dynamic assessments.

Multiagency Enterprise Architecture Planning Framework

Kenneth C. Hoffman, Principal Investigator

Bedford and Washington

Problem
The integration and allocation of information resources and physical assets of multiple agencies is a critical problem in many national programs—homeland security, border control, trade, counter-narcotics, emergency health and safety, and e-government services. The scope and complexity of this environment go well beyond those of single agencies and joint missions.

Objectives
This project will identify the critical planning, strategy, management, and technical activities for complex multi-agency enterprises. It will also identify and evaluate the capabilities of planning tools and methods, both existing and under development, for dealing with these governance and technical complexities. The tools and methods evaluated will be organized in a repository matched to multi-agency governance and technical patterns.

Activities
Research tasks include identifying governance and technical patterns that pose unique planning requirements, activity models and scenarios for missions and services supported by an EA, and next-generation EA and planning methods to support governance and technical objectives. We will also create the Multi-Agency Test Bed and a methods repository matched to multi-agency patterns. This multi-agency scope includes state and local governments, foreign governments, and private sector partners.

Impacts
This research will enable MITRE to select the most appropriate tools and methods and apply them to the full spectrum of multi-agency mission and service support programs. Research results may be incorporated in the Federal CIO Council Guidance (OMB) to deal with the complexity of multi-agency missions and services, and to improve the planning methods and architectures in participating agencies.