The datacentric enterprise an EA pattern

1
The data-centric enterprisean EA pattern

• Natty Gur
• Enterprise architecture group

2
Agenda

• What are EA patterns?
• Pattern definition.
• Business architecture blueprint.
• Information architecture blueprint.
• System architecture blueprint.
• Infrastructure architecture blueprint.
• Governance steps blueprint (lack of time)
• Summary

3
What are design patterns?

• Design patterns proven solutions for well known
  problems.
• Should reference
• Name that describes the pattern
• Problem to be solved by the pattern
• Context or settings, in which problems occurs
• Forces that could influence the problem or its
  solution
• Proposed solution to the problem
• Relations to other patterns and behind the
  solution
• Known uses

4
What are EA patterns?

• EA patterns proven architecture blueprints for
  well defined type of enterprises.
• Should reference
• Name of the pattern
• definitions of a typical enterprise that matches
  the pattern.
• Forces that could influence the match between the
  pattern and enterprise.
• Blueprints of architectures
• Relations to other patterns
• Known uses of the pattern

5
Pattern definitions

• The "data-centric enterprise"
• definitions of the typical enterprise
• Organization for which information is the main
  asset such as Intelligent agencies, IRS,
  Gartner,
• Organization which sees data and data sharing
  across the organization as one of the business
  main concerns.
• Used to be central computer organization (MF)

6
Forces

• CEO changed vision, mission
• Global economic atmosphere
• Acquisitions and merges
• Previous attempts to run EA

7
Business blueprint

• Business processes are integrated and share data.
• Many lane switching in enterprise BPMNs
• BP that gather information from different sources
  into shared repository.
• BP use data intensively
• BP success based on data.
• As a rule of thumb, most users can run any core
  BP.

8
Integrated BPM lanes
Non data-centric
Data-centric
Grid BP
9
Business blueprint

• User should work against one integrated system.
• Need for dynamic and changeable views of
  information.
• Duplication of data or data silos exist, hurt the
  business and should be eliminated

10
Information blueprint

• Information is the atomic element on which the
  enterprise is based.
• Classification of BP data
• Core/administrative
• Core
• Supportive
• Administrative
• Its hard to define clear boundaries.
• Information worlds / By manage users
• Group of data items managed by well defined group
  of workers in the enterprise.
• Those data items managed by the group but
  accessed by other groups.

11
Information blueprint

• Shared/Sector
• Shared data managed by information world but
  being used by other users in the enterprise.
• Sector data managed by information world. These
  are NOT used by any other group.
• Usually every group works with certain raw data
  to produce data that will be shared by others.
• Flower diagram

12
Information blueprint shared/sector
13
Information blueprint

• Information worlds expose
• Visual presentation
• Information world Logic
• Data management (CRUD).
• Information worlds are basic Lego parts for
  building systems
• Cross world indexing/Retrieval
• Each information world should define
• Access principals to information world
• Security
• QoS

14
Information blueprint

• Mapping existing physical schemas to information
  worlds.
• Possibility of historical archive for unused
  data.
• Information worlds solve semantic ambiguities for
  the same notion in different worlds.
• Telephone, Bank account
• Create common enterprise syntax.
• Prevent duplications and silos

15
Systems blueprint

• System as a host for information world.
• System gathering of information worlds with
  workflow.
• System of Systems (SOS).
• Systems developed by different groups in
  different times should be operated as one system,
  from the user point of view.
• Exposing interfaces and services for current and
  future systems.
• Technical dependant
• Technical neutral

16
Systems blueprint

• Using HTML presentation to get
• Integrated systems on one screen.
• Enable swift flow and calling stack while moving
  between systems.
• The same look and feel for all systems
• Personalization of views.
• Basic system structure
• Information world
• Logic/flow
• Expose services
• Consume services

17
Systems blueprint

• ODP systems.
• 8 RM-ODP transparencies.
• Hexagonal over layer
• MVC II for reuse, maintain and infrastructures
• Naming services / proxies
• Application Server
• Basic Interface to implement (by all information
  worlds)
• IsDataContained
• IsDataValid
• ReturnDataByIdentifiers

18
Systems blueprint

• When starting a system you need to define which
  Information worlds should be implemented by your
  system.
• System wont access data directly, just by calling
  the right information world.
• Using the development cycle as part of the
  governance.

19
Infrastructure blueprint

• DB / OODB
• Message queuing/delivering
• Morphology engine.
• Storage Las/Nas
• Backup
• Instant messaging
• Load balancing
• Manage users and permissions AD
• Web servers

20
Infrastructure blueprint

• Indexing/retrieval engine
• GIS server
• Extract knowledge from un-structural documents
• Application servers
• Names extraction engine
• Switches, hubs, firewalls
• EAI
• Document management
• Mail server

21
Infrastructure blueprint
22
Relations

• Do you know other enterprise architecture
  patterns?
• What are the relations?

23
Uses

• At least two governmental bodies in Israel.
• MIT (variation of )
• Do you know others?

24
Summary

• Defined pattern for data centric enterprises
• No perfect match, but might save a lot of time
• Changes and adaptations will be more then
  appreciated.

25
Thank you

• Natty Gur
• E-mail natty_at_theeagroup.net
• Site http//www.theeagroup.net
• The enterprise architects
• 513b Plaza Rd.
• Vestal, NY 13850
• [phone]