Inmon vs. Kimball: Let s get to the bottom of this! Breckenridge Academy Data Warehouse Boot Camp DW Overview, Page 2

DW Boot Camp DW Overview βA reckenridge cademy Breckenridge Academy Inmon vs. Kimball: Let’s get to the bottom of this! TDWI LA Chapter Los Angele...
Author: Diana Rich
11 downloads 2 Views 702KB Size
DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Inmon vs. Kimball: Let’s get to the bottom of this! TDWI LA Chapter Los Angeles, CA March 9, 2010 Speaker: Bob Conway [email protected] 303-885-4811 ©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 2

Anthony’s Pyramid

Strategic Planning

Decision Support

Management Monitoring & Control

Analysis & Reporting

Business Operations

Transaction Processing

Business Perspective

Systems Perspective

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 1

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 3

Why Data Warehousing? Transaction Processing

Analytical/Reporting

•Real-time perspective

•Historic perspective

•Detailed data

•Summarized/derived info.

•Update intensive

•Read-only

•Online updates

•Batch updates

•Batch reporting

•Online reporting

•Performance sensitive

•Flexibility priority

•Structured processes

•Ad-hoc reporting

•Stable data structures

•Evolving data structures

•Functional organization

•Cross-functional

•Clerical community

•Mgmt/analyst community

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 4

OLTP versus Reporting Design Transaction Processing

Analytical Processing

•Highly normalized

•Denormalized

•Minimal indexing

•Liberal indexing

•Transaction logging on

•Transaction logging off

•Record locking on

•Record locking off

•Individual records

•Sets of data

•Calculate derived data

•Store derived data

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 2

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 5

Traditional System Development Purchasing Application

PARTS SUPPLIERS POs

•Data belongs to an application Inventory •Point-to-point interfaces Management

PARTS ASSYs REQUISITIONS Accounts Payable

INVOICES VENDORS PAYMENTS ©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 6

Typical Legacy ‘Architecture’ • Redundant data ($) • Inconsistent, unreliable data ($$$)

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 3

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 7

Data Warehouse Concept Data Warehouse

Reporting and Analysis Read Only Snapshots Historical, Summarized Validated, Integrated

Data Transformation Layer

Extract, Translate Load

Operational Layer (source of record)

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 8

Anthony’s Pyramid •DW complement Tx Proc •DW pathway to integration •DW immediate value

Strategic Planning

DW addresses these

Management Monitoring & Control

Business Operations

Business Perspective

Decision Support

Analysis & Reporting

IE only addresses these

Transaction Processing

Systems Perspective

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 4

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 9

Inmon Architecture circa 1992

Oper DB •op detail •current •App-oriented •Unintegrated

Atomic DW Dept DW •op grain •parochial •latency •summary •subject oriented •derived data •time variant •enterprise integrated

Individual •PC-based •temporary •ad-hoc •heuristic

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 10

Relational (3NF) Design COUNTRY cntry_id

REGION region_id

DIVISION division_id

MARKET market_id

YEAR

PRODUCT_CLASS

ACCOUNT

year_id

prodcut_class_id

account_id

QUARTER INVOICE

quarter_id STATE cntry_id state_id

DEPARTMENT TERRITORY territory_id

department_id division_id

region_id

customer_id market_id postal_cd cntry_id

invoice_id

year_id

CUSTOMER

account_id

PRODUCT_CATEGORY product_category_id

MONTH

prodcut_class_id

month_id quarter_id

COUNTY county_id state_id cntry_id

PAY_PERIOD

WEEK

EMPLOYEE

pay_period_id

week_id

employee_id

PRODUCT product_id

territory_id department_id

product_category_id

month_id

month_id DAY day_id month_id week_id pay_period_id

CITY city_id county_id

ORDER order_id customer_id employee_id day_id

POSTAL_ZONE postal_cd cntry_id city_id

INVOICE_ITEM ORDER_ITEM order_id order_line_id product_id

invoice_line_id invoice_id order_id order_line_id day_id

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 5

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 11

Atomic History Table Design CSTMR_HST cstmr_dwidhst_sqnc_id 101 1 102 1 103 1 101 2

strt_ext_ts Mon night Mon night Tues night Wed night

end_ext_ts** Wed night 12/31/2999 12/31/2999 12/31/2999

hst_crf N Y Y Y

cstmr_nm Bob Joe Mary Robert

101 - Bob 101-Robert Mon Wed 29991231 No gaps or overlaps in time spans for a given DWID value ** Use high date (12/31/2999) instead of NULL for current row

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 12

Inmon Architecture circa 1996

ODS

Oper DB •op detail •current •App-oriented •Unintegrated

Oper Data Store •op grain •latency •subject oriented •volatile •enterprise integrated

EDW Enterprise DW •op grain •latency •subject oriented •time variant •enterprise integrated

Data Marts Dept DW •parochial •summary •derived data

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 6

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 13

Kimball Architecture circa 1996  3NF is great for tx processing but is inappropriate for DW  Dimensional model (star schemas)  Central fact table –aggregate measures  Grouped by dimensions - denormalized

 EDW- collection of star schemas (by subject area) with shared (conforming) dimensions.

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 14

Dimensional (Star Schema) Design D_DAY day_id week_id pay_period_id month_id quarter_id year_id

F_SALES D_EMPLOYEE employee_id territory_id region_id department_id division_id

day_id employee_id product_id customer_id sales_units sales_amount ytd_sales

D_PRODUCT product_id prodcut_class_id product_category_id

D_CUSTOMER customer_id market_id postal_cd city_id county_id state_id cntry_id

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 7

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 15

Proliferation of Data Marts ETL tools Sales

HR

Sales DM

BI Tools

Mfg DM

BI Tools

Finance DM

BI Tools

MRP

Acctg

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 16

DW Challenges in 2010    

Multiple DWs-Redunant/Inconsistent Data Integration/Data Quality Performance/Scalability (ETL and BI) Maintenance/Extensibility  Evolving/Expanding Reporting Needs  Changing/Expanding Source Systems

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 8

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 17

Program Management Office BusReq1

Bus Req2

PMO

StrgcProj1 SP2

SP3

SP4

BusReq3 TactProj11 TP12 TP13 TP14 BusReq4 •Reconcile Business Requirements •Scope DW Projects •Justify DW Projects •Prioritize DW Projects ©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 18

Data Warehouse Concept Data Warehouse

Reporting and Analysis Read Only Snapshots Historical, Summarized Validated, Integrated

Data Transformation Layer

Extract, Translate Load

Operational Layer (source of record)

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 9

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 19

Data Warehouse Concept Order Mgmt SOR1 Extraction Transformation Load (ETL)

Oracle Billing

Data Warehouse SQL Server

SOR2 DB2 •Optimized for Tx processing •Heterogeneous technology •Redundant data •Inconsistent semantics

•Optimized for Analysis Reporting •Integrated data •Detail, Summarized, Historic •Cross-Functional •Enterprise Perspective •Flexibility, Ad-Hoc Access

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 20

Data Warehouse Architecture Order Mgmt SOR1 Oracle Billing SOR2 DB2

Extraction Filter Conform Datatype Conform Domain Error Trapping Consolidate Generate PKs Renormalize Aggregate Track History Load

Data Warehouse SQL Server

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 10

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 21

RAPID Architecture Components

As-Was

SOR1

ETL1

ACQ1

ETL2

As-Is

CNF1 ETL3

SOR2

ETL1

… •Heterogeneous •Redundancy •Inconsistency •Difficult Reports

ACQ2

ETL2

•Persistent •Historic •Non-transformed •Comprehensive* *SOR tables *All rows/cols.

Extract Change Capture Audit Columns

HIST ETL4

BASE ETL5

CNF2

•Transient •Subset •Source Layout* •Target Domains

Filter Tables/Rows/Cols Conform Data Types Conform Values Default/Error

DIM

•Persistent •Integrated •Detailed •Normalized •Non-Historic

Renormalize Match/Merge Gen DW PKs Enforce FK/RI

•Dimensional •Historic •Summarized •Derived Data

Denormalize Aggregation Derivation Time Resolved

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 22

Dimensional Layer Bus. Objects

Cognos

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 11

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 23

DW Layers - Purpose/Usage As-Was SOR1

ETL1

ACQ1

ETL2

As-Is

CNF1 ETL3

SOR2

ETL1

ACQ2



ETL2

HIST ETL4

BASE ETL5

CNF2





• Drill Down • Reusable BASE-DIM • Ad-Hoc Reporting • Operational Reports

• Drill Back • Error Reporting

• Simplify ACQ-BASE • Reusable CNF tables • Expedite Development • Expedite Prod Schedule

DIM

• Simplify Reports • Performance • Std. Derivations • Leverage BI Tools

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 24

Error Tracking/Reporting 1st Logic

RULES

ACQ

ETL2

CNF Inline

ETL3

BASE

Inline ERR Log

BI Tool Offline

Data Stewards

Error Reports ©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 12

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 25

DW Layers - Purpose/Usage As-Was SOR1

ETL1

ACQ1

ETL2

As-Is

CNF1 ETL3

SOR2

ETL1

ACQ2



ETL2

BASE ETL5

CNF2





HIST ETL4

• Drill Down • Reusable BASE-DIM • Ad-Hoc Reports • Operational Reports

• Drill Back • Error Reporting • Retrospective ETL Rules • Simplify ACQ-BASE • Reusable CNF tables • Expedite Development • Expedite Prod Schedule

DIM

• Simplify Reports • Performance • Std. Derivations • Leverage BI Tools

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 26

DW Layers - Purpose/Usage As-Was SOR1

ETL1

ACQ1

ETL2

As-Is

CNF1 ETL3

SOR2



ETL1

ACQ2

ETL2



• Drill Back • Error Reporting • Retrospective ETL Rules • Retrospective DW Scope • Stable ETL1 • Interim Reporting

BASE ETL5

CNF2



HIST ETL4

• Drill Down • Reusable BASE-DIM • Ad-Hoc Reports • Operational Reports

• Simplify ACQ-BASE • Reusable CNF tables • Expedite Development • Expedite Prod Schedule

DIM

• Simplify Reports • Performance • Std. Derivations • Leverage BI Tools

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 13

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 27

ACQ Layer for Interim Reporting SOR1

ETL1

Bus. Objects

ACQ1

V1 SOR2

ETL1

ACQ2

V2 SOR3

ETL1

EII Layer

Cognos

ACQ3

V3 FF4

ETL1

ACQ4

2ndry

Indexes

View Layer

SAS

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 28

Interim Reporting in ACQ Layer  Rapid deployment of ‘tactical’ reports  Offload OLTP reporting  Unified tactical/strategic reporting  Common extraction/delta process/history

 Early insight/experience into the business requirement

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 14

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 29

EDW Architecture Roadmap S1

A1

S2 S3 S4

A2 A3 A4

S1

A1

St1

S2

A2 A3 A4

St2

S1

A1

St1

S2

A2 A3 A4

St2

S3 S4

S3 S4

V1

EII Project 1

V2

EDW Release 1

V3

Hist

EDW Release 2

Base Dim

Project 2

St4

Hist

EDW Release 3

Base

Project 3

Dim

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 30

Re-Acquire Legacy DW Legacy ETL

SOR

ACQ

Legacy Reports

Legacy DW

Legacy ETL

HIS

CNF BAS

SOR …

ACQ …

CNF …

DIM

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 15

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 31

Re-Conform Legacy DW

Legacy ETL

SOR

ACQ

Legacy ETL

Legacy Reports

Legacy DWs

HIS

CNF BAS

SOR …

ACQ …

CNF …

DIM

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 32

Re-Integrate Legacy DW

Legacy ETL

SOR

ACQ

Legacy DWs

Legacy Reports

HST

CNF

Re-Arch Reports

BAS SOR …

ACQ …

CNF …

DIM

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 16

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 33

RAPID Architecture Summary        

Multi-layer, modular design Supports incremental development Durable for expansion Flexible for changes Leverage reusability Strategic and tactical solutions Ad-hoc and structured reporting Technology neutral ©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 34

Questions

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 17

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 35

Bill Inmon on Ralph Kimball The Data Warehouse Toolkit, 1996 Forward: “…Kimball’s stark cognizance and revolutionary approaches…have been tested in crucible of reality. DW Toolkit is one of the definitive books of our industry and mandatory reading for IT professionals…to successfully and profitably conduct business. W.H. Inmon Prism Solutions Aug. 27, 1995

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 36

Retrospective Scope Expansion Release 1

Release 2

Time

Release 3

Release 4

Incremental data elements ©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 18

DW Boot Camp

DW Overview

βA

reckenridge cademy Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 37

ACQ Layer Storage Management XYZ 200 tables •90 SOR •110 Non-SOR

ETL1

ACQ_XYZ 90 SOR •15 In Scope •75 Out Scope

Monthly Archive (Out Scope, CRF=N) (In Scope >120d)

ETL2

CNF_XYZ 15 In Scope

Restore (On-Demand) (Specify tables and dates) Off Line Storage

©Copyright 2010, Breckenridge Academy. All rights reserved.

Breckenridge Academy

Data Warehouse Boot Camp

DW Overview, Page 38

Data Warehouse Critical Success Factors Incremental/ Opportunistic

Technology/ Tools

DW C.S.F.s

Business/IT Partnership

Architecture/ Methodology

Governance Resources/ Organization

©Copyright 2010, Breckenridge Academy. All rights reserved.

©Copyright 2010, Breckenridge Academy. All rights reserved. 19