Composability issues in Network Based Modeling and Simulation

1
Composability issues in Network- Based Modeling
and Simulation

• Farshad Moradi
• farshad.moradi_at_foi.se
• Swedish Defence Research Agency (FOI)
• Dept of Systems Modelling
• Rassul Ayani
• rassul_at_imit.kth.se
• Royal Institute of Technology (KTH), Stockholm

2
Outline

• Network-based Modeling and Simulation (NetSim)
• Motivation, vision, mission
• NetSim architecture, features
• Composability
• Definitions, aspects and challenges
• Composability in software engineering and MS
• Composability in NetSim
• Two approaches
• Future work
• Conclusion

3
Background/Motivation

• Technology driven
• Advances in Web- and Internet-technologies is
  contributing to emergence of a new style of MS
  architecture which is characterized by among
  others component-based methodology with loosely
  coupled simulation models.
• Demand driven
• The concept of Network-Centric Defence, which is
  a SOA, envisions a highly interconnected network
  of what can be generalized as, producers and
  consumers of information. In order to be able to
  integrate MS into future defence systems, there
  is a need for developing MS and MS-related
  services within NCD.

4
Our Vision

• This new architecture suggests development of a
  uniform framework for description and utilization
  of distributed components which facilitates
  reusability, interoperability and composability
  of simulation models.
• The framework should be able to provide MS and
  MS-related services.

5
Our Mission

• To develop a scalable service-oriented
  architecture, which provides a collaborative
  environment for development, execution and
  evaluation of simulation models.
• The environment should facilitate reusability,
  interoperability and composability of simulation
  models through a uniform framework for resource
  description.

6
NetSim features

• Collaborative design, development, and execution
• Distributed resource management system
• Distributed resource repository
• Distributed execution management
• Component-based MS
• Security
• Ability to support different simulation
  architectures
• Toolkit for e.g. model composition, scenario
  development, management, and execution

7
The NetSim Architecture
Overlay Network Services
8
Resource Management

• Distributed Resource Management System, DRMS
• Features of the DRMS
• Storage of simulation components and
  documentation
• Efficient component search and retrieval
• Execution of HLA federations and ...
• Migration of federates
• Load-balancing and Fault tolerance
• Utilization of idle processing capacity

9
Distributed Resource Repository
10
Distributed Resource Repository

• Services
• Distribution of resources
• Up and down loading of files
• Creation and removal of files
• Search
• Semantic search
• Discovery of files and resources
• Security
• SSO (single Sign on)
• Certificates

• Architecture
• Middleware between NetSim-client and the
  underlying architecture
• Carbonara
• Grids, Globus
• GSI (Security service base on PKI)
• RLS (Discovery services containing RLI, LRC)
• GRIDFTP (Up and down loading)
• Metadata, RDF-format
• Different description levels
• Inference engine
• Jena

11
Distributed Execution
12
Distributed Execution
13
Computer-Based Collaboration

• Collaborative MS
• Gives a common picture of the problem, despite
  physical location
• Joining SMEs, developers, VVA people, customers
• Improving and assuring quality of work, enhancing
  efficiency of work, Increasing availability of
  expertise, work process control
• Collaborative Defence MS
• Military Training, Distance Education, Distance
  Planning,Distribution of military competence and
  expertise!
• Collaboration during development and execution
  phases
• Challenges such as, Group Management,
  Synchronisation and Coordination

14
Computer-Based Collaboration Environment
15
Component-based MS
Model (federation) Development
Radar object Radar federate (Aircraft-)
Federation (Air Combat-) Federation
..
.
Component-based Architecture

• Composability, one of the key issues and main
  challenges

16
Composability

• The capability to assemble and combine simulation
  model components in different combinations to
  develop various simulation systems to meet
  specific user requirements. (Petty, 2004)
• Syntactic composability
• Implementation details, such as parameter passing
  mechanisms, external data accesses, and time
  assumptions are compatible for different
  configurations
• Semantic composability
• Can models that make up the composed simulation
  can be meaningfully composed? To make sure that
  the composed model is valid.

17
Semantic and Syntactic Composability

• Examples
• Communication
• Lego

18
Interoperability vs Composability

• The ability of different simulations connected in
  a distributed system to collaboratively simulate
  a common scenario (Petty, 2004)
• Technical Interoperability simulations should be
  compatible with the interoperability protocol
• Substantive Interoperability the exchanged
  information should be semantically meaningful
• Components that are interoperable in one
  configuration and cannot be combined and
  recombined in other way (without significant
  effort) are not composable

19
Composability different aspects

• Factors affecting the difficulty of MS
  composition (Paul Davis, 2004)
• Complexity of system being modeled
• Size, degree of complexity, human factors, etc.
• Difficulty of objective and context
• Uncertainty, rigorous, flexibility, degree of
  plugplay, etc.
• Strength of the relevant science and technology
• Of system, of MS, of management, legacy modules,
  etc.
• Quality of human considerations
• Degree of community, management, human workforce
  capital, etc.

20
Composability different aspects
If management is poor or many technical and
cultural boundaries are crossed or Key science
and military science are poorly understood and
poorly treated
Rigorous applications (e.g., for weapon-system
evaluation)
Risk of failure (despite investment appropriate
to MS size)
Nonrigorous applications (e.g., some training and
discovery experimentation)
Davies 2004
Effective size and complexity
21
Composability in Software Engineering vs. MS

• More similar than different
• Different directions, same results and
  conclusions
• Face the same issues
• CORBA, COM, EJB / DIS, ALSP, HLA

22
Different approaches to composability MS
(Wiesel, 2004)

• Common Library Approach (JMASS)
• Library of reusable software modules
• No stand-alone simulation
• Requires documentation
• Open architecture, simulation development system,
  tools, service, standards and interfaces
• Product Line Approach (OneSAF)
• Contained simulation development system utilising
  layers of products for development of specific
  simulation systems
• Composability through a variety of simulation
  development products
• Services and tools exist to allow dev., config.,
  exec. and analysis

23
Different approaches to composability MS
(Wiesel, 2004)

• Interoperability Protocol Approach (JSIMS)
• Run-time exchange of data or services using an
  interoperability protocol, such as DIS, ALSP or
  HLA (connection through network)
• HLA provides data transfer, semantic comp.
  achieved manually
• Object Model Approach (Base Object Model, BOM)
• Standard for model specification, where models
  are reusable only after modification or the
  development of suitable interface
• Components are not stand-alone simulations
• BOMs to improve interoperability, reuse and
  composability by providing patterns and
  components of simulation interplay to be used as
  building blocks in the assembly of simulations
  and enterprises of simulations

24
Different approaches to composability MS
(Wiesel, 2004)

• Formal Approach (DEVS)
• Depends upon of a simulation formalism to define
  composability in a theoretic and mathematical way
• It is unique in its attempt to prove in a formal
  or mathematical way how models can be composed
• In addition to engineering composability,
  semantic composability theory (Petty, et. al.)
  addresses semantic composability of components
  using a simulation formalism specifically
  designed for that purpose

25
Different approaches to composability Sofware
Engineering
(Bartholet, 2004)

• Predictable Assembly from Certifiable Components
  (PACC) and Prediction-Enabled Component
  Technology (PECT) (Carnegie Mellon
  University/Software Engineering Institute)
• PACC predicts and certifies the run-time
  behaviour of an assembly of components
• PECT is an approach to achieving the PACC
  objectives, an extended component model which
  builds reasoning frameworks, RF, (consisting of
  property theory decision procedure) for any
  run-time property
• RF explicitly exposes any assumptions about the
  system it models
• The component technology then ensures that
  components and assemblies satisfy these
  assumptions through static checking

26
Different approaches to composability Sofware
Engineering
(Bartholet, 2004)

• Web Ontology Language (OWL) and Semantic Web
  (World Wide Web Consortium)
• Defines structured ontologies for delivering
  richer integration and interoperability of data
  among descriptive communities
• Data is described using formal terms including
  discrete math concepts and class hierarchies
• Ontologies are built using OWL to describe the
  structure and meaning of data specific to any
  domain
• OWL also includes formalisms to tie together
  ontologies
• Together with tools for reasoning about
  ontologies could support semantic composability

27
Different approaches to composability Sofware
Engineering
(Bartholet, 2004)

• Unified Modeling Language (UML) and Model Driven
  Architecture (MDA)
• UML is an OMG standard providing a graphical tool
  for modeling the structure, behavior, and
  management of software applications
• MDA, also an OMG product, provides the means to
  separate application logic from platform
  technology
• Platform Independent Model (PIM), describe the
  application while abstracting any potential
  platform technology
• Using standardized mapping with tools the PIM can
  be transformed to Platform Specific Model (PSM)

28
Composability our approaches

• Compsability framework
• A common framework for describing resources
• Common terminology / Common Information Model
• Structured Meta-Models/Component specifications
• OWL, OWL-S
• Open standards
• HLA, XML, etc.
• Library of reusable components
• Model composition environment

29
The NetSim Architecture
Overlay Network Services
30
Composability assumption

• Different parts of the architecture have
  different requirements
• If the meta-models contains sufficient
  information to fulfill those requirements then it
  is enough to guarantee some degree of
  composability

31
NetSim requirements

• Implementation specific information
• Name, programming language, architecture,
  version,
• Run-time information
• Software hardware requirements, initialisation
  arguments,
• Domain specific information/Intended use
• Fidelity, time model, provided functionality,
  simulation type,
• Application interface information
• Information about data that can be exchanged with
  other components
• Semantic information
• Effects, usage restrictions,

32
Composability second approach

• Composability of formal model descriptions
  (formal reasoning) and automatic code generation
• HLA XVCL
• XVCL - XML-based Variant Configuration Language
• NUS
• A method and tool for managing changes during
  software evolution and reuse
• Enforce documentation standards
• Facilitate reuse of common solutions
• Automate any routine but tedious software
  production tasks
• Improve reusability of HLA-federations through
  XVCL

33
Composability second approach

• Similar to MDA
• Develop PIMs and PSMs
• Perform composability check at PIM and PSM level
• Automatically generate executable models
• Repository of reusable PIMs and PSMs

34
NetSim 2004

• Research
• Framework for describing resources (simulation
  models, tools, computing resources, ...)
• Distributed Resource Management (distributed
  resource repository, efficient migration of
  simulation models, fault tolerant executions,
  load-balancing, ...)
• Computer-based Collaboration Synchronization
  and distributed coordination
• Component-based MS
• Distributed network architectures and techniques
• Security issues in NetSim

35
Future work

• Continue work on composability, both approaches
• Experimenting with the meta-models to identify
  sufficient amount of information
• Developing formal description of models, which
  contains semantic information and can be
  converted to program code
• Continued development of NetSim environment

36
Summary and conclusion

• Network-based MS is an important area
• Composability is a key issue
• Semantic composability difficult, but not
  impossible
• Composability issues are the same in software
  engineering and MS. MS could make use of
  achievements in SE
• Two approaches have been taken in NetSim project.
  Both promising, but further development and
  evaluation is required

37
Questions?
farshad.moradi_at_foi.se rassul_at_imit.kth.se