Get Started on SOA - IBM - United States · PDF fileGet Started on SOA 05 ... Run QualityStage...
Transcript of Get Started on SOA - IBM - United States · PDF fileGet Started on SOA 05 ... Run QualityStage...
1
Information Entry Point – Information as a Service
Get Started on SOA
05 - Information as a Service v1.5.ppt 2
Information Management Requirements are Increasing
Disconnected Silos of Information
Dynamically DeliverInformation
Delivering information in context to optimize business processes, applications and productivity
Information Platform
People, Processes, Applications
Silos of Business
Information
70% of people’s time can be spent finding relevant information
60%+ of CEOs say they need to do a better job leveraging information
5X More Value creation by organizations effective at using information
Rich Standards,Flexible Architecture
2
05 - Information as a Service v1.5.ppt 3
Reuse:ServiceCreation
SOA Entry Points
Connectivity:Service
Connectivity
People:Interaction & Collaboration
Services
Process:BusinessProcess
Management
Information:Informationas a Service
SOADesign
SOAGovernance
Service Security, Management and
Virtualization
Getting the right data quickly and consistently for all applications continues to be a key challenge for many enterprises.
Forrester, January 2006
You will waste your investment in SOA unless you have enterpriseinformation that SOA can exploit.
Gartner, March 2005
Service Oriented ArchitectureInformation as a Service is Key
05 - Information as a Service v1.5.ppt 4
The Need for Information Servers
Trust
Productivity
Collaboration
Scalability
Reuse
Today’s Inflection
Point
Federation Delivery SOA
Replication Quality Transformation
Connectivity MetadataProfiling
3
05 - Information as a Service v1.5.ppt 5
IBM Information ServerThe SOA Foundation for Information Management
� A New Kind of Platform� Delivers trusted information to
processes, and applications� Foundation for SOA� Dramatically Simplifies Integration
� Innovation Technology� Metadata-driven integration� Breakthrough productivity� Industry-leading scalability� Rich, bi-directional connectivity
� Based on Experience� 5,000+ Integration Clients� 75 Clients in Beta Program� 35+ Partners Enabled…
IBM Information ServerUnderstand, Cleanse, Transform, Deliver
Trusted Informationin Context
New!
05 - Information as a Service v1.5.ppt 6
Understand Cleanse Transform Federate
Parallel ProcessingRich Connectivity to Applications, Data, and Content
IBM Information Server
QualityStage Federation ServerDataStage
Information Services Director
Metadata Server
Replication ServerData Event Publisher
IBM Information Server
Business GlossaryInformation Analyzer
Data Architect
Delivers Data Access and Data Integration Services for SOA Environments
4
05 - Information as a Service v1.5.ppt 7
Shared ServicesAdministration, Logging, Security, Load Balancing, ….
Integrated Metadata Management
Web Services EJB
IBM WebSphere Information Services Director
Quality Stage Data Stage Federation Server
Enterprise Service Bus
Design Operational
Applications
J2EE Applications
Run QualityStage Job Run DataStage Job Run Federated Query
WISD Agent
WISD Agent
WISD Agent
Process Portal
IBM Information Server – Common Programming Model
IBM Information Server
05 - Information as a Service v1.5.ppt 8
� Foundation for SOA data applications� Common Services infrastructure� Standard access to Common Services
� Services can be invoked as EJB or Web services� Services can be stored in the WebSphere Service Registry and
Repository� Simplified administration and deployment � Unified reporting, logging, and administration
SOA Framework inside IBM Information Server
WebSphere Information Services Director
Shared ServicesAdministration, Logging, Security, Load Balancing, ….
Integrated Metadata Management
IBM WebSphere Information Services Director
Design Operational
J2EE ApplicationsService
Registry and Repository
5
05 - Information as a Service v1.5.ppt 9
Service Oriented Finance Needs a Complete View of the Customer
Service Oriented FinanceCIO
Our Relationship Managers need to be able to see all the customer information at once.
IBM
With IBM’s Information Server you can deliver a federated view of the customer data.
05 - Information as a Service v1.5.ppt 10
Understand Cleanse Transform Federate
Parallel ProcessingRich Connectivity to Applications, Data, and Content
IBM Information Server
QualityStage Federation ServerDataStage
Information Services Director
Metadata Server
Replication ServerData Event Publisher
IBM Information Server
Business GlossaryInformation Analyzer
Data Architect
6
05 - Information as a Service v1.5.ppt 11
Different Types of Integration for Different Needs
� Move data to physical master copy� Large amounts of data moved � Data may require transformation and
cleansing� Typically used for analytical purposes
Master Data
Application
Consolidation(Extract, Transform, Load)
SQL
Sources
Consolidation
Federated Data Access
WebSphereFederation
Server
SQLApplication
Sources
Federation
� Source data remains in place� Small amounts of data accessed
� Typically used for operational purposes
05 - Information as a Service v1.5.ppt 12
Data sources
DB2Checking
Data
DB2 Savings Data
SQL ServerCredit Card
Data
Informix Mortgage
Data
Customer banking data in DB2
Credit card data stored in SQL Server
Mortgage information in Informix IDS
Query
Query
QueryQuery
Multiple queriesMultiple connectionsHarder to program
Problem: Customer Information Is Stored in Multiple Systems
7
05 - Information as a Service v1.5.ppt 13
WebSphere Federation Server
Application ProgramData sources
Checking Data
Savings Data
Credit Card Data
Mortgage Data
SQL Query
WebSphereFederation
Server
DB2 wrappers enable access to a variety of data sources as DB2 virtual tables
Data administrator may then define views across tables
05 - Information as a Service v1.5.ppt 14
DEMO: Federated Data Delivered via IBM Information Server
Data Sources
DB2
SQL SVR
Informix
Create federated query using query designer
Deploy query as a Web Service using IBM Information Server
Service Oriented Finance Portlet uses Web Service to get the federated customer information with a single service request
FederationServer
DataStage
QualityStage
Info
rmat
ion
Serv
erWebSphere Portal Server
Web Service
8
05 - Information as a Service v1.5.ppt 15
IBM Information Server
Checking & Savings
…
Credit Card Data….
Mortgage Data
SQL ServerDB2
Web Service
Portlet
IBM Information Server Benefits
� For the Portlet Developer –� Easily use Web Services to access
federated data� No need to connect to multiple
sources� No need to know various source
schemas� Lab tests show 40-60% reduction in
development time and code size
� For the Administrator –� Easily create Web Services to expose
federated data� Data schemas may be changed
without forcing change to application programs Informix
05 - Information as a Service v1.5.ppt 16
� DB2 UDB � Informix� Oracle� Sybase� Teradata� MS SQL Server� and more …
� Red Brick Warehouse� Classic Federation
� IMS, VSAM, CA-Datacom, CA-IDMS, Adabas
� Ingres� Lotus Notes� Microsoft Access/Excel� MySQL� PostgreSQL� SAS� Sybase SQL Anywhere� and more…
� Web Services� Life Sciences� WBI sources
� SAP, PeopleSoft, Siebel
� Legacy Data Access via WS II Classic Federation for z/OS
� and more …
� Or build custom wrappers� Java and C++ SDK
Non-relational WrappersODBC SourcesRelational Sources
Content Sources
� FileNet� Interwoven� Microsoft � and more …
Sample Data Sources and Wrappers for WebSphere Federation Server
9
05 - Information as a Service v1.5.ppt 17
Understand Cleanse Transform Federate
Parallel ProcessingRich Connectivity to Applications, Data, and Content
IBM Information Server
QualityStage Federation ServerDataStage
Information Services Director
Metadata Server
Replication ServerData Event Publisher
IBM Information Server
Business GlossaryInformation Analyzer
Data Architect
05 - Information as a Service v1.5.ppt 18
Different Types of Integration for Different Needs
� Move data to physical master copy� Large amounts of data moved � Data may require transformation and
cleansing� Typically used for analytical purposes
Master Data
Application
Consolidation(Extract, Transform, Load)
SQL
Sources
Consolidation
Federated Data Access
WebSphereFederation
Server
SQLApplication
Sources
Federation
� Source data remains in place� Small amounts of data accessed
� Typically used for operational purposes
10
05 - Information as a Service v1.5.ppt 19
Service Oriented Finance Needs a Data Warehouse
Service Oriented FinanceCEO
All of the branches are reporting good profitability.
But each branch is reporting the business results differently.
IBM
With a consolidated corporate data warehouse. You can consistently analyze branch performance.
05 - Information as a Service v1.5.ppt 20
Loading your Data Warehouse with WebSphere DataStage
Master Dataor
Data WarehouseExtract
Transform
Load
Multiple databases or files
WebSphere DataStage
SQL ServerBranch Data
SalesFacts
ProductProfitability
Branchs
Product
Star schema tables
Customer Accounts
ProductCampaigns
DB2
11
05 - Information as a Service v1.5.ppt 21
� Create, execute, and monitor jobs to extract, transform, and load data
� Graphically design, develop, deploy, and reuse ETL jobs quickly and easily
� Deploy jobs in real-time, batch mode, or as services using SOAP and/or EJB bindings
� Parallel Processing delivers Unmatched Function, Performance, and Scalability
DataStage®
ExtractTransform
Load
Extract, Transform, and Load (ETL) Data with WebSphere DataStage
05 - Information as a Service v1.5.ppt 22
SALEDATEAMTQTYBRANCHID
CUSTID
PRODID
Data Warehouse
Branch Data
Campaign Sales Data Transformation
AMOUNT DATECUSTNOQTYPRODUCT
Transform
000 101 100 01 01 10,000.00 2007-02-28
000 121 100 01 03 500.50 2007-02-28
000 101 101 01 01 20,000.00 2007-03-01
000 101 01 100 10000 2006-02-28
000 121 03 100 500.50 2006-02-28000 102 01 100 20000 2006-03-31
000 101 100 01 01 10,000.00 2006-02-28000 121 100 01 03 500.50 2006-02-28000 101 101 01 01 20,000.00 2006-03-01
AMOUNT DATECUSTNOQTYPRODUCT
SALEDATEAMTQTYBRANCHID
CUSTID
PRODID
Branch Data
Data Warehouse
Transform
000 101 100 01 01 10,000.00 2007-02-28
000 121 100 01 03 500.50 2007-02-28
000 101 101 01 01 20,000.00 2007-03-01
Different field namesDifferent field order
Add Branch IdentifierDifferent currency format
12
05 - Information as a Service v1.5.ppt 23
DEMO: Run DataStage ETL Job
Branches
DataStage
ExtractTransform
Load
CorporateSQLServer
DataWarehouse
05 - Information as a Service v1.5.ppt 24
DEMO: Use Same Excel Tool to Do Corporate Level Analysis� Excel accesses data from DB2
for z/OS seamlessly
� Campaign to offer mortgage point deduction most effective…Park City branch shows this
� The Inlet branch has the largest number of new accounts for all of the branches, however that does not necessarily equate to the most profit
13
05 - Information as a Service v1.5.ppt 25
WebSphere DataStage Enterprise Editions Provide Real Scalability for Big Jobs
Uniprocessor SMP System MPP, GRID, and Clustered Systems
� Design integration process without concern about data volumes or time constraints
� Leverages database partitioning schemes for optimal load performance
� Simple steps to define partitions within each process if needed
� Single configuration file may be reset to add processors, hardware
� No hand coding of programs to enable more processors
� Supports SMP, Clustered, GRID, and MPP platforms
05 - Information as a Service v1.5.ppt 26
General AccessSequential FileComplex Flat FileFile / Data SetsNamed PipeFTP Compressed / Encoded DataExternal Command CallParallel/wrapped 3rd party appsEMC InfoMoverWeb logsUnstructured: e-mail, docs, etc.Content Management SystemsLife Sciences
Standards & Real Time WebSphere MQJava Messaging Services
(JMS)JavaXML & XSL-TEBXMLWeb Services (SOAP)Enterprise Java Beans (EJB)EDIFIXSWIFTHIPAA
Enterprise ApplicationsJDE/PeopleSoft EnterpriseOneOracle ApplicationsPeopleSoft EnterpriseSASSAP R/3 & BISAP XISiebelJDAAribaManugisticsI2And more…
LegacyAllbase/SQLC-ISAMD-ISAMDatacom/DBDS MumpsEnscribeEssbaseFOCUSIDMS/SQLImageSQLInfomanKSAMM204MS AnalysisNomadNucleusRMS S2000SupraTOTALTurboImageUnifyAnd many more….
RDBMSDB2 (on Z, I, P or X series)OracleInformix (IDS and XPS)IngresMySQLNetezzaProgressRDBRedBrickSQL/DSSQL ServerSybase (ASE & IQ)TeradataUniverseUniDataNonStopSQLAnd more…..
CDC / ReplicationDB2 (on Z, I, P, X series)OracleSQL ServerSybaseInformixIMSVSAMADABASIDMSNonStopSQLEnscribe
IBM Information Server Connects to Almost Everything
14
05 - Information as a Service v1.5.ppt 27
The Need to Improve Data Quality
Critical Problem� Extracted data may have inaccuracies, errors, and
duplicate data which must be corrected to create a “master”consolidation
Why?� No common standards across organization � Human error� Operational data is estimated to degrade in accuracy 2%
per month� Unexpected values stored in fields� No reliable keys for consolidated views
05 - Information as a Service v1.5.ppt 28
Data Cleansing with WebSphere QualityStage
New Customers
1. Standardize Format
2. Match/Compare
to data in the Warehouse
Customers Update
Warehouse
Customers
3a. Duplicate
3b. New Customer
OR
15
05 - Information as a Service v1.5.ppt 29
Service Oriented Finance Needs to Store XML Data
Service Oriented FinanceCIO
We use XML to describe our products.
I need a strategy to store XML data.
IBM
DB2 9 with pureXMLcan manage XML and relational data together
05 - Information as a Service v1.5.ppt 30
Chart Sources: XML Market Opportunities, Forecasts and Strategies, 2004-2009 WinterGreen Research Inc. ZapThink
� More XML data generated than in databases.
� Growing at twice the rate of the total database market. [IDC]
� XML is pervasive in all kinds of organizations.
� Almost every sector has XML-based standards.
Importance of XML?
“XML is emerging as the standard way to represent data for the purposes of interchange between applications and enterprises across
structured and unstructured sources.”—Gartner Group, “Gartner Study on EIM Highlights Early Adopter Trends and Issues,” David Newman, February 7, 2006.
16
05 - Information as a Service v1.5.ppt 31
XML Solves Business Problems Today
� Business-to-Business Integration� Platform-independent transport
mechanism.Transaction orders may be defined in XML
� Forms and Document Processing� Government and legal industry require digital
signatureTax forms require signature & change year to year
� Documents often contain sub-documentsLiterary materials contain books, chapters,
and sub-chapters
� SOA� Web Services messages are XML
05 - Information as a Service v1.5.ppt 32
� Relational is a data model− Relations (tables)− Attributes (columns)− Set based w/ some sequences− Strict schema
� XML is a data model− Hierarchical tree structure− Nodes (elements, attributes,
comments, etc.)− Relationships between nodes− Sequence based w/ some sets− Flexible schema
POID CustomerID ItemID12 1 2162 3 4162 3 5
Id LastName FirstName Street City State Zip
1 Pirahesh Hamid 1 Harry Rd San Jose CA 951413 Selinger Pat 555 Bailey Ave San Jose CA 95141
ItemID Name2 #6 wire nut5 Small Walrus4 Apollo moon rocket
XML – The Difference Is Fundamental
<? xml version=“1.0” ?><paymentRequest id=“12345” >
<payment><payment_name>Loan</payment_name>
</payment><payee_info>
<payee_amt>$97.85</payee_amt></payee_info><remitter>
<name>John Smith Co</name><address>
<street>1234 W. Main</street><city>Yonkers</city><state>NY</state><zip>11111</zip>
</address></remitter>
</paymentRequest>
17
05 - Information as a Service v1.5.ppt 33
� Simplified XML Data Access� Centrally store and access XML data� SQL or XQuery can be used to retrieve XML data� Join XML data with it’s related relational data
� Search Speed� Search XML documents quickly and efficiently
using search optimization engines
� XML Data Needs Protection� Transactional protection� Backup and recovery to ensure continuity� Security mechanisms for access control
� Optimize Existing Investments� Use existing technology infrastructure and skills to
store and manage both relational and XML
����
XML Data Management Needs Relational MaturityWhat Is Needed?
05 - Information as a Service v1.5.ppt 34
Solution: DB2 9 Native XML Storage� A database server which supports both the relational and XML data
structures� Adds a new “XML” data type
� A new storage mechanism to efficiently manage XML data� “Native“ means that XML documents are stored on database pages
as parsed tree structures to reflect XML’s hierarchical structure� Separate from the relational storage mechanism
� Native XML support avoids conversions between XML and relationalstructures, and the corresponding limitations. � Input and retrieval are faster, performance is better, and querying is
better and faster.� With conversion techniques like BLOBs and shredding, every
operation (parsing, etc.) is expensive and there is a potential loss of data.
� The XML document might be too complex to shred.
18
05 - Information as a Service v1.5.ppt 35
DB2 XML Performance Beats Oracle
XML document Insert Performance
2.2
0.0
0.5
1.0
1.5
2.0
2.5
DB2 VS Oracle 10G Structured XML storage
x Fa
ster
DB2Oracle
XML document Mixed Read/Update/Delete Performance
1.5
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
DB2 vs. Oracle 10G Structured XML storage
x Fa
ster
DB2Oracle
XML document Retriaval using multiple Read queries Performance
3.6
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
4.0
DB2 VS Oracle 10G Structured XML storage
x Fa
ster
DB2Oracle
DB2 9 performs 1.5 to 3.6 times faster than Oracle 10G
on XML operations
05 - Information as a Service v1.5.ppt 36
60% (3x better)18%ORDERS
Compression Ratio
59% (2x better)29%Entire Database
58% (1.5x better)38%LINEITEM
DB2OracleTable
� TPC-H is a well known data warehouse benchmark� Each vendor uses the same tables and same data� Oracle published their compression rates for TPC-H tables at the
VLDB conference in 2003� IBM ran the same tests on the same tables
� Test results – DB2 reduces cost by requiring less storage
DB2 – Better Compression Ratio Than Oracle
19
05 - Information as a Service v1.5.ppt 37
Challenge
Solution
Benefits
� Reduce development time and improve response time when using XML data
� Reduce development time and improve response time associated with storing XML
� Schema changes in minutes not days� New database search and retrieval
methods using native XML in minutes…
Implications
� Deeply integrated XML support enables unique client application opportunity
� XML data services critical to SOA implementations
� DB2 9 a combined XML / Relational Information Server
Storebrand Leverages Integrated XML Support
05 - Information as a Service v1.5.ppt 38
What About Our Competitors?
NONOYES8. Native XML data
LimitedNOBEST7. Data Compression
**LimitedLimited
connectivityLimited
Basic ODBCYES1. Federated database support
NONOYES2. Information Server platform
LimitedNOYES6. Data cleansing and harmonization
LimitedMultiple products
NOYES5. Highly scalable ETL - uniprocessor to true grid
LimitedSQL transforms
LimitedSQL transforms
YES4. Graphical tool for automatic ETL job development with built-in transformations functions
NONOYES3. Cross-product metadata repository
OracleMicrosoftIBM
20
05 - Information as a Service v1.5.ppt 39