Presentation of Patterns in Data Quality and MIT IQIS

1. Proceedings of the Second
MIT Information Quality
Industry Symposium
Massachusetts Institute of Technology
Cambridge, MA, USA
July 16-17, 2008
Edited by
WooYoung Chung
St. John Fisher College, USA
Suzanne Acar
U.S. Department of the Interior, USA
Linda Kresl
Mentor Graphics, USA

2. Copyright © 2008 MIT Information Quality Program, Massachusetts Institute of
Technology
All presentations included in these proceedings are published with explicit permission from their
author or copyrights holder. These presentations are protected by the international and U.S.
copyright law. Any copying, distribution, exhibition in public or private meetings, and/or
pursuing any derivative work, by any third party, in part or as a whole, is strictly prohibited
without explicit written permission of the original copyrights holder. Any third party who wishes
to quote or refer to the information presented in these proceedings must comply with the fair use
provisions of the U.S. copyright law 17 U.S.C. §107 and must properly cite the author and these
proceedings as its source.

3. FOREWORD
WELCOME to the 2nd MIT Industry Symposium on Information Quality
(MIT2008IQIS). Information Quality (IQ) is an acknowledged prerequisite to
organizational success. However, organizations also face the challenge of balancing
quality requirements including an increasing volume of information and the demand for
faster delivery, while contending with constantly changing business environments. The
Symposium is designed to help bridge these issues through discussions among
practitioners, vendors, and academicians. In addition to presentations and workshops, the
Symposium includes vendor presentations, product announcements, and consultancy
methods that complement the annual International Conference on Information Quality
(ICIQ).
Acknowledgments are extended to all Symposium participants. We thank you for
your contribution in establishing IQ as a multi-disciplinary field and industry. Members
of the Symposium organizing committee worked hard; their contributions made our work
on the program a pleasure. We also wish to thank Lockheed Martin Corporation, UTi
Worldwide, and Acxiom Corporation for their continuing sponsorship of the MIT IQ
Consortium.
Thanks are also due to, among others, DAMA, EWSolutions and DM Forum for
their support in promoting the Symposium.
At the MIT 2007 Information Quality Industry Symposium, Dr. Bruce Davidson
(Cedars-Sinai Health System) and CDR Stanley Dobbs (U.S. Navy) were recognized for
their outstanding contributions.
We would like to express our gratitude to Fori Wang, Li-Hsin Chang, and other
MIT staff for their assistance in producing the Symposium proceedings, keeping the
website up and managing the various aspects of the Symposium operation. The final
Symposium program and other information about the Symposium are available at
http://mitiq.mit.edu/IQIS.
i

4. MIT Information Quality Industry Symposium, July 16-17, 2008
The MIT 2008 Information Quality Industry Symposium
Patterns in Data Quality
A Method for Organizing
Enterprise Data Quality (Web) Services
in Service Oriented Architectures
Michael Overturf – VP of Strategy
Navin Sharma – Dir. of Product Management
The MIT 2008 Information Quality Industry Symposium
Patterns in Data Quality
• A methodical approach to structuring rules for data
user satisfaction
• Patterns provide simplification
• Patterns focus measurement
• System architects are the primary beneficiaries of
simplification
179

5. MIT Information Quality Industry Symposium, July 16-17, 2008
The MIT 2008 Information Quality Industry Symposium
Background
Structured constraint patterns developed by Christopher Alexander
(Architect and Mathematician) in 1964 in ‘Notes on the Synthesis of
Form’
M – Set of Design Constraints
L – Set of Design Constrain Relationships
G(M,L) – Linear Graph of Design Constraints
Design Constraint Design Constraints
Relationships
The MIT 2008 Information Quality Industry Symposium
Constraint Tree Dynamics
Definition-Inheritance dynamic describes how we get from DQ
Constraint Set to DQ Service
Definition Inheritance
Set Template
Class Composite
Types Atomic
180

6. MIT Information Quality Industry Symposium, July 16-17, 2008
The MIT 2008 Information Quality Industry Symposium
Constraint Patterns in Data Quality
Data quality constraints
define
the form of data quality rules
Metadata Platform Access Modality Input Output
Consistency Operating System Web Transactional Type Type
Services/SOAP
Completeness SaaS C/S, API Batch Binding Binding
Accuracy Security Microbatch Source Sink
Uniqueness/ Multi-modal (Any)
Singularity
The MIT 2008 Information Quality Industry Symposium
Pattern yields Form Set
A simple example
Class
Types
Designer
Atomic
Composite
Template
181

7. MIT Information Quality Industry Symposium, July 16-17, 2008
The MIT 2008 Information Quality Industry Symposium
Pattern yields Form
A simple example
Subflow
Free Input Deployed Free Output
Parser
The MIT 2008 Information Quality Industry Symposium
Quality as a Service (QaaS) Template
Composite
Patterns naturally translate into
services (in a SOA context):
Atomic
– Atomic services that are fine grained
to provide a monotonic function;
– Composite services that encompass
two or more atomic services;
– Templates make up one or many
composite and atomic services
by specific data quality metrics
(consistency, uniqueness, etc.)
industry specific business processes
182

8. MIT Information Quality Industry Symposium, July 16-17, 2008
The MIT 2008 Information Quality Industry Symposium
Example:
Business
Insurance
Process Underwriting
Template
Composite
Service
Atomic Services
The MIT 2008 Information Quality Industry Symposium
Web Services implement Pattern Constraints
• Widely adopted and accepted within enterprises and vendor
communities
• Software community agreement on standards for distribution
– WS-I and W(3)C standards,
– WSDL (Web Services Description Language)
• Principles encourage and enable –
– Ease of integration (loose coupling)
– Re-use
– Ease of management
– Agility
– Active use
• Satisfies modality constraints (batch, transactional)
183

9. MIT Information Quality Industry Symposium, July 16-17, 2008
The MIT 2008 Information Quality Industry Symposium
Laying the foundation Integrate
• Create – a Composition Orchestrate
Interface that allows data
stewards and enterprise Develop
architects to use and define
templates, composite and
atomic data quality services Secure
• Publish – for distribution
Analyze
• Integrate – via standards
based interfaces (WSDL, etc.)
• Manage – for governance Access Manage
Consistent Measurement of Data Quality Results
The MIT 2008 Information Quality Industry Symposium
Summary
• A methodical approach to structuring rules for
data user satisfaction
• Patterns provide simplification
• Patterns provide a structure for measurement of
data quality
• System architects manage data quality using
standard Web Service Management Lifecycle
184

10. MIT Information Quality Industry Symposium, July 16-17, 2008
The MIT 2008 Information Quality Industry Symposium
Contact Information
• Michael Overturf
– Email: Michael_Overturf@g1.com
– Phone:413-695-5500
• Navin Sharma
– Email: Navin_Sharma@g1.com
– Phone: 240-447-6801
185