Title: Challenges in Automating the
1Challenges in Automating the Provisioning of
Parametric Initialization Data to Simulation
Applications
Lee Lacy, Mike Puhlmann, Gabriel Aviles, Wayne
Randolph - Dynamics Research Corporation Jin
Kwon AMSAA Major Matt Chesney - U.S. Army
TRAC-Monterey
2Background
- AR 5-11 Management of Army Models and
Simulations - Share valid data to all MS data consumers
- Develop Standards to use common data
- Minimize Cost of Data
- Higher resolution model support
- Data management
- State of the practice
- Data interchange formats
3DIF Implementation Challenges and Potential
Solutions
- Semantic Interoperability,
- Semantic Mapping Responsibility,
- Explicit Tags vs. Meta-model Approach,
- Standard Nomenclature,
- Entity Type Enumerations,
- Versioning / Traceability,
- Storage Methods,
- Distribution Methods, and
- Standards Development Process
4Semantic Interoperability
- Challenge although XML can help solve syntactic
interoperability challenges, differences in
producer and consumer semantics (the meaning of
the data) must be addressed in other ways. - Proposed solution standardizing on data models
and providing explicit semantics.
5Semantic Mapping Responsibility
- Challenge although a DIF provides a standard
data format, both producers and consumers will
likely require a mapping process to translate
their data to/from their data models into the
DIFs semantics. - Proposed solution decisions must be made
regarding whether to delegate transformation
requirements to the producer or the consumer.
6Explicit Tags vs. Meta-model Approach
- Challenge
- Explicit DIFs use tag names that contain the
identifier of the data value being passed - The meta-model approach involves using a single
tag name and passing the data value identifiers
as text string data - Proposed Solution Compromise embed explicit
tags only when necessary in a Meta Model
7Standard Nomenclature
- Challenge common naming is a significant, yet
easily solved challenge in sharing simulation
data. A variety of schemes are available - Proposed solution armys standard nomenclature
database (SND) for equipment and munition naming
used by army analytical community in support of
army studies and the DMSO common semantics and
syntax effort for other parametric descriptions
8Entity Type Enumerations
- Challenge assignment of unique identifiers to
simulation object types - Proposed solution Modernized Integrated Data
Base (MIDB) over the IEEE Distributed Interactive
Simulation Enumeration
9Versioning / Traceability
- Challenge pedigree or provenance of the data is
especially important to verification and
validation agents - Proposed Solution provide metadata with the data
that indicates the version and pedigree of the
data. The metadata may be needed down to the
individual data items level
10Storage Methods
- Challenge large file sizes and classification
levels - Proposed solution Physical storage not a huge
limiter but using short tag names, using tabs
instead of spaces, limiting embedded comments,
using explicit DIFs, and subdividing documents
will help
11Distribution Methods
- Challenge distributing data
- Proposed solution
- Immediate solution is to use web portals that use
human intervention. - The future vision is to enable web services that
automatically respond to consuming software
requests for data. The goal should be to
decrease the amount of human intervention.
Included services should be able to check the
version of data and update the data where needed
12Standardization Process
- Challenge standards Interoperability ontology
and other agreements must be developed in a
collaborative environment with input from various
interests and compromises on the approach. - Proposed solution SISO and other industry
groups develop and document standards. Develop
Recommended Practice Document
As defined by AR 5-11, Data Standards is A
capability that increases information sharing
effectiveness by establishing standardization of
data elements, data base construction,
accessibility procedures, system communication,
data maintenance and control.
13Summary
- Semantic interoperability
- Semantic mapping responsibility
- Explicit tags vs. Meta-model approach,
- Standard nomenclature,
- Entity type enumerations,
- Versioning / Traceability,
- Storage methods,
- Distribution methods, and
- Standards development process
- Provide explicit semantics
- Decision Consumer or Producer
- Compromise combined data model
- SND and DMSO common semantics
- MIDB before IEEE
- Provide archive info meta data
- Limit tag names, spaces and comments
- Web Services
- SISO
14Backup Slides
15Providing CP Data with a DIF
Export to common XML format
Import to simulations formats
Data Consumers (e.g., OneSAF)
XML DIF
Authoritative Data Sources in Multiple Formats
Simulation System Data in Multiple Formats
16Data Mapping
17Summary
18Questions?