C2IEDM Data Mediation Service

1
C2IEDM Data Mediation Service

• Bo Sun, Andreas Tolk
• Virginia Modeling Analysis Simulation Center
• Old Dominion University
• Norfolk, Virginia, USA
• bsun,atolk_at_odu.edu

2
Outline

• Motivation
• What do I do with my data models in case of
  C2IEDM mandates?
• Data Mapping
• Preparation
• Documentation
• VV
• Implementation
• Model Based Data Management
• Implementing Data Mediation Services

3
Motivation

• Why should I Care?
• What do I do with my data models in case of
  C2IEDM mandates?

4
Why care about Data Mediation

• Global Information Grid (GIG) and other
  Service-oriented Architectures are likely to
  become the backbone of future distributed
  applications
• Joint Command and Control
• Future Combat Systems
• NATO Allied Command Control System
• General Challenge Composable Services
• How do I ensure that I understand other services
• How do I ensure that other services understand me

5
How do I migrate my components

• Rewrite everything using a common model
• Cost prohibited
• Most applications have task-tailored views
• Enterprise-wide data model not feasible
• Establish necessary connections to other
  services
• Point-to-point solutions (2n challenge)
• Only feasible for existing solutions
• Leaves interpretation to the implementer
• Data Mediation Services
• Based on Common Understanding

6
Data Mediation Service Vision
7
Current SolutionDDMS Primary Category Sets
DoD Discovery Metadata Specification
(DDMS) National Security Information Sharing
Standards for Resource Metadata Application
Profile for Discovery (NSISS RMAPD)
8
Data Mapping

• Preparation Phase
• Documentation Phase
• VV Phase
• Implementation Phase

9
Data Mapping

• Data Mapping is the Conceptual Mapping.
• Mapping Complexity depends on relational database
  complexity of both source and target data model.
• Data Mapping includes four phases

Data
Validation and Verification
Implementation
Documentation
Preparation
10
Preparation Phase

• Use a Consistent Method for source and target
• IDEF1X
• Content attributes
• Structure attributes (keys, foreign keys)
• Identify Business Objects
• Analyze foreign keys
• Identify exchange dependencies
• Identify first mappings on Conceptual Level
• Similar properties and concepts

11
Documentation Phase

• Document the conceptual mappings
• Top-down mapping
• Document additional properties needed
• Bottom-up mapping
• Document the resulting key dependencies
• Business-object to business-object mapping
• Applicable tools
• Excel
• Property and propertied concept matrix

12
MS Excel
13
Property and Propertied Concept Matrix
14
Data Validation Verification

• Validation
• Am I doing the correct thing?
• SME (operational/tactical) necessary
• Verification
• Am I doing things correctly
• Are all targeted attributes covered?
• Are all targeted concepts covered?
• Are the functions correct?
• Documentation
• Can another expert understand what I did and why

15
Implementation

• Objective
• Software Layer translating source into target on
  the fly
• Applicable tools
• Extensible Stylesheet Language Transformation
  (XSLT) tools
• Additional Function Layers for
• Aggregation
• Disaggregation
• Conditional transformations
• etc.

16
Model Based Data Management

• Common Reference Data Models
• Data Engineering

17
Components of Data Engineering

• Data Administration
• Managing the information exchange needs incl.
  source, format, context of validity, fidelity,
  and credibility
• Data Management
• Planning, organizing and managing of data, define
  and standardize the meaning of data as of their
  relations

• Data Alignment
• Ensuring that data to be exchanged exist in all
  participating systems
• Data Transformation
• Technical process of mapping information elements
  to each other (often implemented in gateways and
  interfaces)

18
Web-based Standards supporting Data Engineering

• XML as the common syntax and format of all
  components
• Data source registers data description following
  the idea of Universal description, discovery, and
  integration registries (UDDI)
• Mapping of data will be management of tag sets
• After data management using tag set, data
  alignment becomes one-to-one comparison
• Data management can lead to XSLT schema for data
  translation

Potential for Automation of Data Administration,
Data Alignment and Data Translation based on Data
Management
19
Use Common Reference Model

• Use C2IEDM as target and source
• Over 1,500 data elements in app. 200 tables
  representing 176 agreed operational IER
• If data element is in this pool
• Map data element to C2IEDM
• Else
• Extend C2IEDM (standard necessary)
• Map data element to new element

20
Model-Based Data Management
21
Implementing Data Mediation Services

• Tools
• Web Services
• Feasibility Study

22
Implementation Technologies

• Architecture
• Data Mapping/Data Translation
• Web Service
• Software packages
• MapForce Data Mapping
• Glue Web Service

23
VMASC Reference Implementation

• Advantages
• Efficient
• MapForce is Code Generator
• Glue has Web Service Frame
• Effective
• MapForce is Visual Mapper
• Flexible
• Build a new mediation service in no time at all

24
Mapping Exampleusing MapForce
25
Summary

• Data Mediation cannot be done automatically
• To much SME knowledge necessary
• AI applications can only support
• Common Reference Model is necessary
• Common Namespace
• Gradual alignment
• Common Standards necessary
• Endorse C2IEDM as core solution
• Develop and standardize C2IEDM extension rules