多层次的数据设计
Requirements, Process Models
Input - Achieve alignment between business and IT
Conceptual Data Models
Context and concepts, independent of technical consideration
Logical Data Models
Takes architecture into account (Relational, OO, XML, etc.)
Physical Data Models
Takes Platform into account (Sybase, Oracle, MySQL, etc.)
Sybase Confidential Proprietary. 7
概念数据模型 (CDM)
The Business Architect’ View
s
No regard to storage architecture or technology
Map Concepts and Relationships Between Them
Establish Standards
Business naming standards
Sybase Confidential Proprietary. 8
概念数据模型 (CDM)
Support business rules, domains, shared data
items, inheritances and relationships.
A CDM can generate LDMs/PDMs/OOMs.
Notations : Entity/Relationship, Merise, E/R+Merise,
IDEF1X, Barker
Sybase Confidential Proprietary. 9
Barker Notation
Facilitate customer migration from
Oracle/Designer
Use Composite structure to define subtypes
Provide an option to switch from/to Barker
Notation
Sybase Confidential Proprietary. 10
Barker Notation
CDM # ID Long integer
Party
(Barker) Organisation
* Organisation Name Characters (64)
r o Organisation Description Text
Order Person
# Order Id Long integer * Person Name Characters (64)
o Initials Characters (5)
r2
Item
Line Item # Item ID Long integer
# Item Line Integer r3
Product Item
* Product Name Characters (128)
Service Item
Service Type
r4 * Srvice Name Characters (128)
# Service T ype ID Long integer
Sybase Confidential Proprietary. 11
Content of a CDM
Entities –An entity represents an object about which you want to store
information
Example: product, customer, sale
Attribute – a piece of information tracked for each entity.
Example: Customer –name, address, phone
Relationships
Binary Relationships (1:M, M:M, 1:1)
Supertype/Subtype Relationships
Business Rules
Terms, Facts, Derivations, Constraints, more…
For example, "The start date should be inferior to the end date
of a project."
Sybase Confidential Proprietary. 12
Entities and Attributes in CDM
Entities collect a series of Attributes
Attributes are assigned Data Items
Data Items are Unique Cells of Data Definitions
Glossary of Business Terms
Achieve Consistency of Definition at Business Level
Data Items are:
Name, Description
Data Type, Domain
Attributes can be Primary Identifiers
Sybase Confidential Proprietary. 13
数据类型和域
Data Type Specifies the form of the data
corresponding to the domain, such as Numeric,
Date, Boolean, or others.
Domains help you identify the types of information
in your model. They define the set of values for
which a column/entity attribute is valid.
Sybase Confidential Proprietary. 14
逻辑数据模型(LDM)
A logical data model (LDM) helps you analyze the
structure of an information system, independent of
any specific physical database implementation.
An LDM is less abstract than a conceptual data
model (CDM), but does not allow you to model
views, indexes and other elements that are
available in the more concrete physical data model
(PDM).
Notations : Entity/Relationship, IDEF1X, Barker
Sybase Confidential Proprietary. 15
逻辑数据模型 (LDM)
The Designer’ View –Scoped to Technology
s
Relational, XML, OORDBMS
Reflection of Business Requirements
Validates technical Interpretation of Business Requirements
Basis for Physical Implementation
Based on CDM –Generation to/from CDM models
Sybase Confidential Proprietary. 16
Content of an LDM
Logical Entities
Complete Attribute Sets (fully attributed)
Everything needed to manage storage and integrity of data
Attributes Wholly Owned by Entities (no Data Items)
Logical Relationships
Must resolve Many to Many (too much ambiguity)
Primary Identifiers and Foreign Keys
Key Migration and Joins Defined
Normalized
No redundant data, focus on form not RDBMS performance
Sybase Confidential Proprietary. 17
物理数据模型 (PDM)
Database Builder’ View –RDBMS Platform Specific
s
Denormalized for RDBMS efficiency
Storage I/O capabilities
RDBMS-specific capabilities
Redundant Data May be Used to Improve Performance
Derived from the LDM or directly from a CDM
Bi-directional synchronization from CDM:LDM to PDM
Notations: Conceptual, IDEF 1/x Notations, Codasyl
Sybase Confidential Proprietary. 18
Content of a PDM
Tables, Columns
Keys (PK, AK, FK)
Indexes and Constraints
Triggers and Procedures
Domains and Data Types
Storage Details
Tablespaces, Segments, Partitions, etc…
Security Details
Users, Groups, Rights, Permissions
Everything RDBMS
Sybase Confidential Proprietary. 19
物理数据模型 (PDM)
Supports denormalization (split table by rows or by
columns, merge tables, copy columns) for
optimizing speed
Generate and reverse engineer all major databases
Synchronize database with PDM
Estimate database size
Test data generation
View data
Sybase Confidential Proprietary. 20
物理数据模型 (PDM)
Data warehouse modeling
Star schema, snowflake schema
Multi-dimension diagram
Generate Cubes for SQL Server
Generate Oracle Cubes (11.0)
……
Import/export many formats using Meta Integration
Model Bridge
Import Erwin models
Sybase Confidential Proprietary. 21
信息生命周期管理
Information lifecycle modeling allows you to model
the movement of data from expensive, rapid storage,
through various forms of cheaper slower storage as
the data ages and access requirements diminish.
Sybase Confidential Proprietary. 22
信息生命周期管理建模
PowerDesigner provides a simple modeling
structure to cost effectively manage “aging”of
data inside the data center
PowerDesigner help generate scripts to automate
the creation, movement, and purging of partitions
across your lifecycle phases.
Sybase Confidential Proprietary. 23
信息流模型(ILM)
PowerDesigner Information Liquidity Model (ILM) is
a graphical modeling environment focused on data
movement and data transformation.
You can analyze and document where your data
originates, where it moves to, and how it is
transformed on the way, including replications and
ETL.
Sybase Confidential Proprietary. 24
信息流模型(lLM)
Model ETL and EII data transformations and
Replications.
Generate and reverse engineer Replication
Server/MobiLink.
Generate automatic PowerDesigner reports to
document your model.
Sybase Confidential Proprietary. 25
Diagrams of ILM Model
Information liquidity diagram
Sybase Confidential Proprietary. 26
Diagrams of ILM Model
Data transformation diagram
Sybase Confidential Proprietary. 27
Diagrams of ILM Model
Transformation control flow diagram
Sybase Confidential Proprietary. 28
Metadata Repository and Impact
/Lineage Analysis
The metadata repository can store and manage
versions for all the metadata
Source database schema, target database schema,
replication, synchronization, data warehouse and data mart
design, …
Impact Analysis –to analyze the effect of the action
on the objects that depend on the initial object.
Lineage Analysis –to identify the objects that
influence the initial object.
Sybase Confidential Proprietary. 29
0 comments
Post a comment