Master Data Integration Standards
Version
2.0
TRIM file number
-
Short description
Describes the standards to which data being shared or
copied between one or more application must conform
to. All data transferred between applications must be
handled by the middleware layer (webMethods)
Relevant to
Officers who have a responsibility in the planning,
analysis, design, or implementation of a Master Data
Integration (MDI)
Potential suppliers and developers of applications
Authority
Responsible officer
Enterprise Architect - Integration
Responsible office
Enterprise Architecture & Liaison, Division of Information
Technology
Date introduced
January 2010
Date(s) modified
July 2011
Next scheduled review date
To be determined.
Related University documents
CSU Enterprise Architecture Principles
CSU Application Standards
CSU Identity Standards
CSU IT Infrastructure Standards
CSU Security Standards
webMethods version 8.2 adapter documentation
Related legislation
Key words
Integration, principles, standards, Integrations
architecture, webMethods, rules, master data, shared
data
Enterprise Architecture
Master Data Integration Standard
Version 2.0
Description
The CSU concept and principle of “One University” has strong and clear correlation with the integration principles and standards. As
such the integration standards in this document are designed to ensure all CSU applications are able to share amongst themselves
information in a standardised format
By Master Data Integration (MDI) we mean using data abstraction through master data definitions to provide a unified interface to
facilitate uniform data access for viewing and utilising data.
Protocols
Data transferred between internally hosted and externally hosted systems as well as 2 externally hosted systems is only
permissible if transport layer supports https or FTPS via the SSL or SFTP. This
Master Data Caching
Middleware Suite: WebMethods version 8.2
When a new MDD/EII is created, a copy must be provisioned in the MDC in order to provide an alternate read only source of
the definition for web services to use. Data structures (tables) in the MDC must reflect the pure MDD they correspond to.
Integration Methods (In order of preference)
Multi-Mastering
1.
2.
3.
4.
5.
Utilising the Services Oriented Architecture (SOA) by invocation of a web service (W3C compliant)
Publish / Subscribe to database tables
Flat file Transfers via WebMethods
Access to data via a Directory Service (only when hosted on CSU infrastructure)
Other Point to Point methods requiring explicit written approval from EA&L
a. Direct Vendor or CSU developed point to point integrations
b. Direct SQL access to Master Data via MDC
Coding Methods
All integrations designs using methods 1,2 or 3 must access data through the CSU defined native document structure and
NOT using direct SQL from within a Flow service.
Application Development and Acquisition
It is a requirement of all applications to be able to support at least one of the above methods for both consuming and
contributing data in the EII process. Applications that are unable to utilise one of the first 4 standards may be reviewed at regular
intervals and an alternative solution that offers a more sustainable integration solution may be adopted, particularly if the
application is or has been identified as a origin data host.
Master Data Formats
Data must only be transferred from the middleware layer (webMethods) and the destination system in a format which
enforces constraints of the appropriate MDD. Once each Master Data record has passed through the middleware layer and been
validated by it may then be remapped to the required destination system data format (i.e. Transformed Master Data)
Hosting Models
Internally as well as externally hosted solutions are supported by webMethods. Consideration should be given the potential
latency involved with using an externally hosted solution as a provider of data for a web service. A tolerance of no more than X
milliseconds is seen as acceptable for propagation of data, which is separate to the execution time of the service.
Connectivity
Where an application has no platform independent API for webMethods to interact with, priority is given to direct access to
platforms supported in the Infrastructure standards document. The 2 database platforms that have the highest degree of support
are the current versions of Oracle and Microsoft SQL server which are specified in the Infrastructure Standards document. Should
webMethods require direct connection to an operating system, the platforms CSU have expertise in are also stated in the
Infrastructure standards document
Multi-mastering of MDDs is allowed. Multi-Mastered data must be communicated amongst peers via MDD constructs
(webMethods Canonicals) to ensure standardised communication between masters and eliminate any complexities associated with
point to point style mastering
Directory Services / Authentication / Authorisation
The 2 supported directory services for the integration of applications are Active Directory (AD) and LDAP. Whilst their
primary focus is to facilitate authentication, attributes considered core to authentication/authorisation process may be accessible to
applications as a form of basic integration. They should also only be used to integrate systems hosted on CSU infrastructure
Business Continuity / Disaster recovery
Any origin or destination system’s business continuity measures must remain transparent to webMethods. Only in extreme
DR cases will a webMethods configuration be altered to accommodate a change necessitated by a disaster recovery plan being
enacted.
Exceptions
As the level diversity within the application portfolio increases at a rapid pace, new applications requiring EII may exhibit certain
behaviour which may not be explicitly covered in the integration standards. If this is the case it should be raised in an EA&L
invitational meeting where a decision will be made to either incorporate the EII requirements into the integration standards to
encompass new technologies and/or methodologies which adhere to the integration principles or grant a one off exception. In the
event the standards cannot be modified to accommodate an application requiring EII, then an exception to the integration standards
can be considered on a case by case basis.
Exceptions, either partial or total to particular standards may be considered in the event that
1.
2.
A compelling business case is presented to EA&L
Data involved is classified as Level 1 in the Information standards.
Master Data Integration Glossary
Term
Master Data Definition (MDD)
Origin Data Attribute
Master Data Dictionary
Master Data Domain
Origin Data
Recognised Data
Local Data
Replicated Master Data
Transformed Master Data
Master Data Cache (MDC)
Data Custodian
Master Data Catalogue
Origin System
Destination System
Enterprise Data
Enterprise Information Hub (EIH)
Brief Definition
Universally agreed data structure to facilitate the sharing of data
A field of data of which many can be combined to for a master data
definition
Catalogue of all master data definitions
Grouping of master data definitions within an area of activity within
the organisation
Data identified in one or more systems that populate master data
definitions
Data identified in one or more systems that populate master data
definitions but have yet to be implemented.
Data in a system that has not been identified as for use in a Master
Data Definition
A populated master data definition stored in a location different to
its source and maintains MD Definition structure
A populated master data definition stored in a location different to
its source and maps MD definition to destination system data
structure (custom client view of MDD)
Stored copy of every populated master data definition, independent
of any application
A person or division which is accountable for a collection of source
master data attributes
Summary list of all master data definitions
(business view)
System which holds authoritative source data
System which receives copy of master data
Every original data element stored within every information system
within the university (not copies of master data)
Collection of master data definitions, origin data, MDC and
collection of services to access and manipulate master data