1. DEVS-Based Simulation Web Services for Net-Centric T&E Saurabh Mittal, Ph.D. Jose L. Risco-Martin*, Ph.D. Bernard P. Zeigler, Ph.D. Arizona Center for Integrative M&S, Univ. of Arizona *Departamneto de Arquitectura de Computadors y Automatica, Madrid, Spain

2. Outline Need for Net-centric M&S paradigm Position of DEVS in addressing capabilities Components of Net-centric M&S Underlying technologies DEVS, XML, Web Services, DEVSML Distributed simulation using SOADEVS SOADEVS Client Demonstration Conclusions and Future work

3. Need for Net-centric M&S Paradigm Acute need for a new testing paradigm that could provide answers to several challenges in a 3-tier structure Tier 1: individual systems or programs Tier 2: SoSs in which interoperability is critical Tier 3: Enterprise level, where joint and collaborative operations are conducted The paradigm has to comply with Department of Defense Architecture Framework (DoDAF). Unfortunately, DoDAF and other DoD net-centric mandates post significant challenges to current testing and evaluation tools

4. State of DEVS Technology Desired M&S Capability for T&E Solutions Provided by DEVS Technology Support of DoDAF need for executable architectures using M&S DEVS Unified Process [31] provides methodology and SOA infrastructure for integrated development and testing, extending DoDAF views Interoperability and cross- platform M&S using GIG/SOA Layered Simulation architecture for interoperability testing Net-centric composition and integration of DEVS ‘validated’ models using Simulation Web Services Automated test generation and deployment in distributed simulation Model and Simulator separated Bifurcated Test and Development process Test artifact continuity and traceability through phases of system development Model-Continuity principles Real time observation and control of test environment Variable-structure component modeling to enable control and reconfiguration of simulation on the fly Dynamic simulation tuning, interoperability testing and benchmarking.

5. Components of Net-Centric M&S Every web-based simulation consists of: The Application Model Partitioner Model Deployer Model Initializer Model Simulator Many examples exist that implement above steps but in different manner: DEVS/P2P, DEVS/RMI, DEVS/CORBA Most are focused towards simulation engine Our effort is more oriented towards application layers i.e. the modeling application itself

6. Underlying Technologies DEVS Based on mathematical formalism using system theoretic principles Separation of Model, Simulator and Experimental Frame Atomic and Coupled types Hierarchical design Web Services using XML Service Oriented Architecture (SOA) consists of various W3C standards Machine-to-machine interoperable interaction over the network based on WSDL interface descriptions Client server framework Message encapsulated in SOAP wrapper which is in XML

7. Underlying Technologies …contd’ DEVSML Built on JavaML, which is XML representation of Java Behavior representation takes its power from underlying JavaML

8. DEVSML and SOADEVS

9. Distributed Simulation using SOADEVS Fundamental Objective: Integrate software application as services Two layer service framework User layer Upload, Compile, Simulate (centralized or distributed) Engine layer Initialize, DEVS-protocol relation services, exit, console output retrieval service Engine Layer: Initialize simulator i Run Transition in simulator i Run Lamda function in simulator i Inject message to simulator i Get tN from simulator i Get ta from simulator i Get consoleLog from all simulators Finalize simulation service

10. Distributed Simulation using SOADEVS … contd’

11. Abstraction of Coupled Model with an Atomic Model having DEVS State Machine

12. Distributed Simulation using SOADEVS … contd’ Centralized Simulation

13. Distributed Simulation using SOADEVS … contd’ Real-time Distributed Simulation Involves direct coupling with machines

14. SOADEVS Client

15. A Demonstration DEVS Unified Process and SOADEVS

16. Conclusions A SOA-oriented framework called SOADEVS for test and evaluation of DEVS models Centralized and Distributed executions Modified real-time DEVS simulation protocol to be tailored for SOA Layered service architecture with underlying helper services Part of the larger DEVS Unified Process (DUNIP) Communication using XML as middleware A means towards an end

17. Future Work Net-ready capability testing requirements Execution of DEVS test-models in tandem with real web services over SOA Multi-platform simulation capability DEVSJAVA and DEVS/C++ together in a single simulation session interfaced using WSDL specifications of layered SOADEVS Platform Specific Messages (PSMsg) translated to Platform Independent Messages (PIMsg) at run-time using DEVSML architecture