1. Building the Business Case for Addressing Semantics in Application Integration
Eric Lambert (EDF R&D)
Greg Robinson (Xtensible Solutions)

2. Andre, Jean-Luc, Eric, and Greg
Bonjour!

3. Impacts of Deregulation on Business Processes
EDF adapted some of its former BP
EDF created new ones
EDF has to anticipate (i.e : DER on MV network and impact on BP, …)
Information System require more agility
More integration points have appeared
But : 50% of system integration costs are attributed to semantic issues
the 50% of system integration cost is of common acceptance : Gartner Group studies …

4. A solution through a pragmatic use of standards
Adoption of a Model Driven Integration :
Reduces labor to maintain overlapping data in multiples applications
Systematically generated common structure and common vocabulary reduces design time effort as well as coding errors
Reduces performance errors caused by inconsistent information.
Support for managing Key Performance Indicators (KPIs) and handling regulatory obligations.
Performs faster implementation of application functionality and business processes.
Reduces cost to maintain and extend existing applications.
Reduces risk of project schedule and budget overruns

5. International Context
UCA : User groups
TC57
Open Application Group
W3C
MultiSpeak
(NRECA)
WG9 Distribution Feeders
Coordination WG19
WG7 Control Centers
CIM/61850
WG17 ?
WG18
WG14
DMS
WGs 10
Substations
EPRI
UCA2
Project
WG13
EMS
WG16
EPRI
CCAPI
Project
ETSO : European Transmission System Operator : defines standards for European TSO. Relies on UN-CEFACT standardisation process EFET : federates the traders in Europe
ebIX : electronic business for Information exchange => more oriented in the Distribution Area
OLE Process
Control
(OPC)
Object
Mgmt.
Group
ebXML
OASIS

6. Perceived and real problems with standards
The CIM (Common Information Model) is still evolving
CIM is in English (only)
“Establishing a common language” is an unreachable utopia
A methodology deriving XML message types from a UML model is not necessary
Compliant products are not offered yet on a large scale in the marketplace
CIM is only in english : it is a strong need to have the ability to translate CIM definitions in other languages. And CIM definitions already in english should be defined in accordance to ISO11179 (which means no ambiguity) How can we resolve the dilemna : as CIM still evolves, compliant products can’t be offered

7. Answers to some of these problems
The intended use of the CIM is for inter-application integration, not intra-application integration
The CIM is aimed at being a technical integration language
The model is requisite for ensuring that every element of every message is used consistently across messages on an enterprise-wide basis
Several rounds of interoperability tests have been and will be performed
CIM User Group has been created

8. EDF R&D Cimergy Project Key numbers
Started as an innovative action in 2003
Labeled as an EDF R&D project in 2004
Budget : 1,5 M USD / year
Team ~ 7 people
A innovative action is transformed in a project if it is seen promising. Mid 2006 the project will have to answer to the question : « Based on more than 3 years of work,  does CIM is or is not an opportunity for EDF operational Divisions (Generation, Transmission, Distribution, Supplier, ..) ? » In order to facilitate the adoption of the standards by Operational Divisions, the project has to provide a methodology and tools

9. Goals of Cimergy
Define methodology and tools requirements in order to use IEC TC57 standards 61970,
Participate actively in the TC57 standardization effort
Promote the methodology and standards inside EDF
Understand and help the harmonization process
IEC TC /61968 and 61850
IEC TC57 and UN-CEFACT, ETSO, ebIx
Reduce the gap between IT people and Automation people
Is or is not
the CIM an opportunity for EDF Operational Divisions ?
A key issue has been to solve the « problem » : CIM is seen in Europe as a north American standard, not used in Europe. In europe ETSO is well known and is defining standards. We will see that the methodology approach combining IEC CIM Model, and UN-CEFACT Core Component Technical Specification is the answer to this kind of « argument » Reduce gap between IT people & Automation people can be expanded to adress a more global problem : How I can facilitates communication between teams inside a company, how I can federate people around a common goal, how I can avoid
Island of automation ? As an EDF R&D project, the main objective is to help operational divisions, and EDF group to take clear decisions about CIM standard. What are the risks to go or not in that direction, and why ?

10. CIM Approaches
Bottom-up approach : 2 applications need to exchange Data, CIM is used for specifying interfaces
Several Transmission/Distribution applications have been wrapped with a CIM import/export interface
Top-Down Approach : Main Objective of the project in 2005
Model Driven Integration approach
Based on the UN-CEFACT Modeling Methodology (called Core Components Technical Specification), CIM Model, and ISO 11179
Uses cases completed with Distribution Division
CIM Products experimented : MDI (Xtensible Solutions)
2 approaches because we need to be pragmatic, and it has to be proven that the standard can be used, and that a return on investment exist. For concrete API realized see next slide.

11. CIM based API for several functions
CIM Distribution API
GIS
CIM
XML
LV Planification function
MV functions (Load calculation, …)
CIM API for EDF Products
In order to illustrate bottom Up approcah. GIS Export CIM data to several Distribution functions in a CIM based format. The GIS is Smallworld.
CIM API have also been realized on CIM EDF products which are provided at the international level (Prao in Distribution, Eurostag at transmission level). In that case, the API are based on profiles already defined as CPSM (NERC profile, based on CIM), either on CDPSM for Distribution. The key idea of CDPSM is to leverage on what has been done at transmission level with CPSM.
Prao and Eurostag were able to demonstrate a load flow calculation based on CIM data files imported. A wrapper had to be written in order to translate this common format to their proprietary format.
These APIs have been demonstrated during 2004, 2005 interoperability tests organized by EPRI.
PRAO (MV Planification function)
CDPSM Profile
CPSM Profile
EUROSTAG (Power System Dynamics)

12. ‘Integrating’ the Integration Infrastructure
Complex & Evolving
Industry Standards,
Methodologies &
Architecture Patterns
Evolving Utility
Business Execution
Requirements
Evolving
Technologies &
Products
Evolving
Business
Partnerships
The
Model Driven
Integration
Process

13. Methodology : Combining CIM & UN-CEFACT Core Components
UML Information
Model
UML Exchange Model
UML Contextual Model
XML Information Model
XML Contextual Model
XML Exchange Model
The UN/CEFACT modeling methodology states that Core Components in every Business Domain have to be defined. In the Electrical field, IEC is responsible for standardization. CIM is an Electrical information model and can provide Core Components in the UN/CEFACT CCTS : for instance a circuit breaker, a Power Transformer, … A MoU exist between UN/ECE, IEC, ITU, ISO which states that harmonization has to take place between these organisations, and that every normalisation entity is responsible in its field. IEC is responsible of the Electrotechnical Business Field.
This pictures describe all the levels of the methdology ans highlights the UML fiel don the left and the XML field on the right. So far tools have limitations in order to derive a complete methodolgy.

14. A closer look… CIM XML World
Constraints on Association + Assembling rules
XML
XSD
Generic CIM
Message Schema generation
Reprocessed
Constraints on attributes
Here is a closer look, from the core component view, to a contextual view, which constraints the CC view. Same approach for deriving the Message view from the Contextual view

15. Working With Business Participants
EDF Group Information System Division
Generation Information System Department
Transmission Information System Department
Distribution Information System Department
Supplier Information System Department
Project
Information & Telecommunication Division
EDF R&D Division
Working with Business Participants implies many actors. This pictures tries to summirize the global organization and the communication streams that exist between these entities around Business Information Systems Some decisions are made at the EDF group level, and are based on return of experience from operational divisions, information & Telecommunicaition division, and r&d division.
Each operational Division has a certain degree of automnomy, but a consistent framework is used at several levels in order to manage complexity, reduce operational costs, anticipate on promising technological trends.

16. Successfully Scaling Up
Having a service mindset
Working with business units
Working with their (business unit) partners
Proactive change management (next slides)

17. Resistance to Change Human resistance my be the biggest hurdle
25% will like the change & look forward to it
25% will hate change
50% will wait and see
The neutral zone
Between the way things were and the way things will be (during the project)
Marked by confusion and uncertainty
No clear markers and no promises

18. Forms of resistance
Almost everyone has concern about measuring up in a new environment
Resistance can take many forms
Constant questioning
Forms of confusion
Silence or easy acceptance

19. Change Management Driving Forces Restraining Forces Status Quo
Opportunity to learn new skills
Availability of training and tools
Lack of training/understanding
Power of internal expert
Inertia – why change?
Feeling that the job may be threatened
Not invented here
Our problems are special
Status Quo
[Source: Douglas K. Barry]

20. Overcoming Resistance
Anticipate resistance in advance, before the project begins
Select the right people
Start by identifying the right kind of skills and experience
What is not available internally must be obtained externally either through hiring or contracting
A big factor in failed projects is a lack of personnel with the skills and experience required
Pairing team members together
People in pairs should not have the same issues

21. Overcoming Resistance
Really listen
Sometimes the person is only voicing symptoms, not the main problem
Communicate at many levels
Do not over-promise and then not meet the promise; it’s sets a foundation of mistrust
Seek appropriate avenues to involve people
Ask for people’s input and review
Get resistance out in the open
Talk about it in a neutral and non-threatening way

22. Adopting Standard Data Element Definitions
Driving Forces
Restraining Forces
Easier exchange of data
Reduced development time
Reduced maintenance costs
Costs to develop standard definitions
Costs to change existing systems
Existing data definitions are different
Some definitions need to be different
Products use different data definitions
Lack of industry standard definitions
Mergers and acquisitions
Status Quo
[Source: Douglas K. Barry]

23. Adopting A Service-Oriented Architecture
Driving Forces
Restraining Forces
Interoperable networked applications
Emerging industry-wide standards
Easier access to enterprise-wide data
Easier exchange of data
Consistent enterprise-wide data
Reduced brittleness using tags
Support of Web Services in products
Reduced development time
Reduced maintenance costs
Availability of external services
Minimal effect of operational systems
Use of business intelligence software
Availability of training and tools
Opportunity to learn new skills
Mergers and acquisitions
Cost of development
Product/service doesn’t do everything
Deciding what data to route
Delays getting data updates distributed
Deciding what data to warehouse
Delays in getting data to the warehouse
Redundancy of data
Data quality issues
Effect on operation systems for up-to-the
moment data requests
Lack of training/understanding
Power of internal “experts”
Inertia – why change?
Feeling that job may be threatened
Not invented here
Our problems are special
Business
Issues
Design
Issues
Status Quo
Change
Issues
[Source: Douglas K. Barry]

24. Conclusion
Combining CIM model and UN-CEFACT Core Component Technical Specification is a promising approach
The reusable methodology provide end-to-end requirements traceability
As more projects leverage the infrastructure, more data becomes available as part of one coherent body of information
Faster integration of application systems and information should improve EDF’s ability to react quickly to business changes
Less dependence on individual vendors
There is one methodology and managed tool set to ensure consistency and leverage
A key issue is to standardized the methodology in order to get products implementing this methodology.