Integration Patterns for Virtual MDM

Implementations Bhavani Kumar Eshwar

Senior Technical Staff Member
CTO – MDM Engineering, India Software Labs

Coverage – MDM Technology Upgrades, Portfolio

Integrations & Cloud/Bluemix/DataWorks Offerings

© 2015 IBM Corporation

Please Note …
IBM’s statements regarding its plans, directions, and intent are subject to
change or withdrawal without notice at IBM’s sole discretion.
Information regarding potential future products is intended to outline our general
product direction and it should not be relied on in making a purchasing decision.
The information mentioned regarding potential future products is not a
commitment, promise, or legal obligation to deliver any material, code or
functionality. Information about potential future products may not be
incorporated into any contract. The development, release, and timing of any
future features or functionality described for our products remains at our sole
discretion.

Performance is based on measurements and projections using standard IBM

benchmarks in a controlled environment. The actual throughput or performance
that any user will experience will vary depending upon many factors, including
considerations such as the amount of multiprogramming in the user’s job
stream, the I/O configuration, the storage configuration, and the workload
processed. Therefore, no assurance can be given that an individual user will
achieve results similar to those stated here.
2 © 2015 IBM Corporation
 Agenda

1. Recap – Virtues of Virtual MDM

2. Integration Patterns Overview

3. Capabilities associated with the

Integration Patterns

4. Closing Notes

3 © 2015 IBM Corporation

Virtues of Virtual MDM – A recap
Virtual MDM Virtual MDM Virtual MDM
Core facets Benefits What InfoSphere MDM SE offers
•Also called Registry Style MDM •A virtual MDM style can facilitate a •Virtual master registry to assemble a
rapid, cost-effective implementation trusted, tailored view
•Source systems author and own the
data •Does not require changes to source •Workflow capabilities to implement
systems policies and processes for data
•In most cases the Master Data is not stewardship and data governance
modified •Serves best to Organizations in the
early stages of a Unified Governance •Flexible and extensible data model
•Organizations can match/link strategy built for MDM
duplicate records and produce
golden views •Complete, highly accurate, real-time •Highly accurate probabilistic
view of master data thoroughly matching and search
•Golden Views but created on the fly supported by configuration
i.e. current point in time •Policy management to author,
•Offers quick time-to-value monitor and enforce data quality

•Solutions framework to enable and

enhance applications and processes
with master data

•Data stewardship to inspect and

resolve data quality issues

•Relationship and hierarchy

management to manage household
and B2B relationships

•Last but not the least – Key

Integration capabilities

4 © 2015 IBM Corporation

 Integration Patterns for Virtual MDM Implementations

INTEGRATION PATTERNS
OVERVIEW

5 © 2015 IBM Corporation

Three key patterns to look at …

Salesforce

Billing
Acquire Manage DW
Deliver

Patterns for Data Acquisition Patterns for managing the Patterns for delivering data
and Ingestion Virtual MDM hub from Virtual MDM hub

6 © 2015 IBM Corporation

 Integration Patterns for Virtual MDM Implementations

DATA ACQUISITION AND

INGESTION

7 © 2015 IBM Corporation

Integration Patterns – Data Acquisition and Ingestion, the big picture
 The first stepping stone to success in Virtual MDM implementations is fluent and extensible methods for
Data Onboarding from Sources
Acquire
Data Sources Extract/Transform Curate Transfer Load

Deal with all kinds data Employing methods to Landing prepared data Execution of Data
Early DQ subroutines
sources prepare data for MDM onboarding onboarding flows

On-Premise • Data Analysis

Sources
• Full data
extractions

• New Source
addition
SaaS Applications
• Optional in • Using ETL Tool • ETL Job
• Transformation
Virtual MDM
• Landing • Message
• Used for Messages for processing
Standardization MDM to
• Delta Loads • Add/Update
only for better process
matching type Client API
• Messaging
• Using Client calls
XaaS APIs
• Events • Sometimes
used for • Add/Update
deterministic • Using Services Web Service or
de-dupe per Rest API calls
Other Cloud
Hosted / Bluemix source • Using Change
( CDC can be rolled into
Applications • Data Model Data Capture
ETL Jobs )
Change

• Security & Tech

considerations

ERP Bluemix • Compliance

considerations

8 © 2015 IBM Corporation

Data Acquisition and Ingestion - Patterns & capabilities

 Items to cover for the background

 Core interfaces

 Data Analysis ( external to Virtual MDM )

 Loading data into Virtual MDM in a Batch ( applicable for Initial Loads and Batch Loads )
 ETL Tools
 Change Data Capture

 Processing source systems messages

 Processing source system events

9 © 2015 IBM Corporation

Core Interfaces

 Virtual MDM offerings typically provide certain core interface packages for –
 Managing data in the Virtual MDM hub
 Performing certain actions like Searches etc.
 Extension points to perform additional processing in regular out-of-box flows

 InfoSphere MDM SE has –

 SDK for Java and Web Services as the Default Client API Interfaces
 Enterprise Service Oriented Architecture (SOA) Toolkit
 Message Broker Suite
 Event Notifications
 Handlers

10 © 2015 IBM Corporation

Data Analysis - Information Analyzer Integration
Use Cases
• For use in inspection of sources

• Data Validation and metrics for

source and trusted Data

• Source data Profiling to fix

initial Data issues

• Also aides in provide key inputs

for Matching configurations

• Sample asset available on IBM Web Management ( IWM ) site

• The sample includes an InfoSphere Information Analyzer project referred to
as the StartUp IA Project

• This project includes a set of pre-defined rules that are specific to

InfoSphere MDM and the start-up data model

• The IBM InfoSphere Metadata Asset Manager (IMAM) tool is used to import
the schema metadata into the InfoSphere Information Server common
repository

• InfoSphere Information Analyzer uses the common repository when the

data analysis is performed. Data rules and column analysis are run on the
InfoSphere Information Server

• The results are analyzed using the reporting capabilities of InfoSphere

Information Server
11 © 2015 IBM Corporation
Data Analysis – Information Analyzer Integration, Sample Attribute Frequency
Report

MDM Analysis Sample Rules

• Attribute Frequency
What percentage (%) of the Last Names are from source A, source B, and so
forth

• Attribute Completeness
How many Last Name fields are populated, how Last Name fields have no
value (NULLs), and so forth

• Attribute Validity
This includes passing in a set of anonymous (ANON) values and checking
which ones are ANONs
Validating with an external source
Validating reference data

• Attribute conformance
Data of birth (DOB) falls within a set range, such as from 1910 to 2013
Format checking in general, such as Drivers License (DL) always use digits
12 © 2015 IBM Corporation
Loading data into Virtual MDM in a Batch

Source Data Extraction Virtual MDM bulk

(full or delta ) loading solution
If data preparation requires
Call backs and complex
transformations for multi-
row payloads

Using Change Data

Change Data Capture Capture to prepare Virtual MDM bulk
from Source Systems payload loading solution

Virtual MDM Inbound

Messaging Processing

If data preparation can be

covered with simple
mappings

13 © 2015 IBM Corporation

Virtual MDM Bulk loading solution Virtual MDM bulk
loading solution

 MDM Connector Stage

 Mpxdata Utility and InfoSphere DataStage

 Legacy Component - Clover ETL

14 © 2015 IBM Corporation

Virtual MDM Bulk loading solution – MDM Connector Stage
 A native Stage in Information Server which can be used in DataStage and QualityStage Jobs
 Currently supports functions to load data into or extract data from InfoSphere MDM SE
 Uses MDM MetaData provided by MDM Workbench’s MetaData Export Capability
MDM Development tasks Metadata Integration Tasks DataStage/Quality Stage
Job Development Tasks
DataStage/
MDM IMAM Asset QualityStage
Workbench Manager Designer

Exports the project Import XMI files, Configure MDM

metadata as XMI analyze, preview and Stage with MDM Hub
files for use by MDM upload to Metadata Connection and other
Connector Server settings

MDM Model Bridge

Query MDM
Model in
XMeta

XMI Files
Persist Stage
XMeta configuration in
SCM or (MDM, ASCL)
XMeta
DSX Model
DevOps (DSX)
Repository
Compile,
[ MDM Design MetaData ] Deploy & Test
Jobs [ Job MetaData ]
15 © 2015 IBM Corporation
Points to note about MDM Connector Stage

 MetaData Export cover InfoSphere MDM SE and InfoSphere MDM AE

 Data Reads and Writes of MDM Connector Stage currently support InfoSphere MDM SE
and Healthcare offering which run on InfoSphere MDM SE
 memget and memput calls are currently supported
 Implementations need variations and enhancements for the read/write capabilities
served by the memget and memput API callouts and this should be available in a future
release

 On the Road map

 memsearch, memscore, memdelete, tskget, tskput, tsksearch etc.
 MDM SE Engine batch processing functions like Batch Compositor and Bulk Cross
Match
 Support for other Editions

16 © 2015 IBM Corporation

Virtual MDM Bulk loading solution - Mpxdata Utility and InfoSphere
DataStage
 Some implementations still leverage mpxdata utility for Data Ingestion into InfoSphere MDM SE
 MEMCOMPUTE mode in conjunction with madhubload and/or madentload utilities
 MEMPUT mode

 The challenges around this approach is that there is no standards for Data Acquisition from sources which also elevates
the problem of end-to-end automation of the Data Acquisition and Ingestion flows

 A sample InfoSphere DataStage Job is provided to assist in designing this “end-to-end” automation of the Data Acquisition
and Ingestion flows

MDM
Workbench
MDM SE
Metadata

IMAM Import
Job
Template
InfoSphere Information Server InfoSphere MDM SE
XML
2. Invoke utility on
Copy to filesystem MDM SE Server JMX Job
Utility
mpxdata run job

1. Drop prepared
source extracts

17
Sources
© 2015 IBM Corporation
Data Acquisition and Ingestion - Processing source systems’
messages and Events

 InfoSphere MDM SE provides the Message Broker Suite

 Has various options for inbound and outbound message processing
 Incoming messages from sources systems either in the form of events or payload
messages can be processed by the Message Broker Suite to Ingest data into MDM
Standard edition

 On the road map - IBM Integration Bus

 Good set of our customer are consolidating to have Reliable Messaging component be
IBM Integration Bus
 Multiple requests for Message Broker Suite capabilities to be made available in IBM
Integration Bus
 Current plan -> Technical Preview by November 2015 and OOTB capabilities in 2016

18 © 2015 IBM Corporation

 Integration Patterns for Virtual MDM Implementations

MANAGING THE VIRTUAL MDM

HUB

19 © 2015 IBM Corporation

Integration Patterns – Managing the Virtual MDM Hub, the big picture
 Managing a Virtual MDM Hub entails many functions some of which are specific to the core Engine and
some deal with integrations
Manage
Enrichment Auditing,
Matching &
& Remediation Monitoring Events
Linking
Validations & Reports

Working with External Handling Key events

Required for Golden Managing adherence Virtual MDM Hub’s
Data providers and provided by the Virtual
Views to Rules and Policies Status & Reports
other integrations MDM Hub

 Most of the topics around Managing the Virtual MDM Hub relate to Governance
 Integration with IBM’s Business Process Manager, IBM Operational Decision Manager,
Cognos etc. will fall under this area
 These is a full session to cover this area
 https://events.na.collabserv.com/portal/wippages/register.php?id=c248284145&l=en-US
• Master Data Stewardship & Governance by Jay Limburn - 30 Jun 2015, 10-11AM EST
 The key topics we will cover under this section are –
 Federated Query across Virtual MDM Hubs
 Integration with InfoSphere Information Server’s QualityStage
 Specialized Integrations for Free Text Searches – IBM’s Watson Explorer Integration
 Working with Events from InfoSphere MDM SE
20 © 2015 IBM Corporation
Managing the Virtual MDM Hub – QualityStage Integration

 Provides for Address Standardization via a default pre-interaction handler

 Customer handler too can be written

AttributeMaps/ pre-
interaction Handler InfoSphere MDM SE

MDM Configuration
Workbench
AddressStdHandler
Callback deploy
OR
Handler Custom Handler

InfoSphere Information Server

Job’s Service Interface

Japanese
Address Std. Name and
Job Address Std.
Job

https://www.youtube.com/watch?v=T5jlMjOrxsQ
21 © 2015 IBM Corporation
Managing the Virtual MDM Hub - Federated Queries

 Provides cross-
organization identity
governance without a
centralized database

 Ideal Information
Exchange and cross-
jurisdiction public sector
use cases

22 © 2015 IBM Corporation

Managing the Virtual MDM Hub – Flexible Search and Watson
Explorer connector
 Flexible Search
 Provides search capabilities for members without relying on bucketing in InfoSphere
MDM SE
 Attributes to be indexed are configurable
 Free Text and Free Term Searches are enabled
(IxnMemTextSearch and IxnMemTermSearch interactions)
 Java & Rest APIs are provided
 Member data for the configured attributes is pushed into the Fuzzy Search component
used i.e. IBM Watson Explorer

 IBM Watson Explorer Connector for InfoSphere MDM SE

 IBM Watson Explorer also ships with a connector for InfoSphere MDM SE
 Crawls thru InfoSphere MDM SE repository
• Member Type Name, Entity Type Name and optionally Composite View Name is
used as configuration for crawling over
• Payload can be calibrated using Segment Code Filters
 Redbook : Building 360-Degree Information Applications
• http://www.redbooks.ibm.com/abstracts/sg248133.html?Open
23 © 2015 IBM Corporation
 Enhanced 360º view of the customer
Janet Robertson Customer search:
MDM draws in all related records:
J Robertson, Janet Robertson,
Jan Baker

MDM enables a complete

purchase history, including
Jan Baker’s records from
before 2011

Customer’s
Products
from MDM

Customer info
from MDM

Unstructured internal
information related to
customer
Indexed 3rd party
information related
to customer

24 © 2015 IBM Corporation

Managing the Virtual MDM Hub - Working with Events from InfoSphere
MDM SE
 InfoSphere MDM SE produces events based on certain conditions while processing
workloads

 These event’s obviously have implications and the key to those are the ones which may
entail Stewardship actions or delivery of the output of the associated views to consuming
systems

 Will be covered in the next section, where we look into Patterns for delivering data from
Virtual MDM hub

 Some of the key Events produced by InfoSphere MDM SE

Entity Created Member Process Task Created
Entity Deleted Member Created Task Deleted
Entity Updated Member Deleted Task Resolved
Member Linked Task Updated
Member Merged
Member Reprocessed
Member Undeleted
Member Unlinked
Member UnMerged
Member Updated
Member No-link
25 © 2015 IBM Corporation
 Integration Patterns for Virtual MDM Implementations

DELIVERING DATA FROM THE

VIRTUAL MDM HUB

26 © 2015 IBM Corporation

Integration Patterns – Delivering data from the Virtual MDM Hub, the big picture
Deliver
Full & Delta Data End-point
Real-time Data Access Events
Extracts Transformation delivery

APIs providing real- Handling Key events Getting data out of the Dropping off data to
Managing adherence various end-points
time access to golden provided by the Virtual Virtual MDM Hub
to Rules and Policies
views MDM Hub based on requirements

On-Premise
End points
Warehouses
Big Data
Environments

SaaS Applications
• Using Client • Processing • Using ETL Tool • Using ETL
APIs MDM Events to Tool
notify end • Selective
• Using Services points or Extracts • Pre-
routing processing
• Using Flex • Full Extracts data
Search • Routing MDM
Events to ETL • Mapping to
XaaS
Tools End-points

Other Cloud
Hosted / Bluemix
Applications

Bluemix
ERP

27 © 2015 IBM Corporation

Delivering Data from the Virtual MDM Hub - Patterns & capabilities

 Real-time Data Access

 Covered in Core Interfaces
 APIs to Search/Get Data

 Handling Events
 OutBound Brokers
 Custom Message Processing
 Hooking Queues to ETL for Processing

 ETL Usage – Virtual MDM Bulk Extract Solution

 Triggers: Event, Full Extracts, IDs for Selective Extract
 Data Collations and Transformation and delivery to end points
 MDM Connector Stage
 Legacy Component - Clover ETL

28 © 2015 IBM Corporation

 Virtual MDM Bulk Extract solution – MDM Connector Stage

InfoSphere Information Server Configurations

Implementation Specific Job trigger Composite View Output
Output
link
(Event or Full Extract of Delta Extract ) Output
link
links

Columns to
Output Link
Map

Row-sets Output Any outbound

Link processing Stages …
Stage with Implementation
specific MDM Keys sourcing File
Output
strategy Link PureData for
Analytics
Hash key (data) values, Range, Row-sets
Round robin, Random, Entire, Output Salesforce.com
Modulus, Tiered database Link
(memget API) Business Applications
partitioning, Dynamic
Output BigInsights etc.
Link

MDM Metadata Information Server’s MDM

Interface components

Information Server
Meta Data Repository

MDM Standard Edition

Instance
(Person, Patient, Provider,
Clinical Data Hub) Master Data Extract Sample Jobs
are provided

MDM
29 Database © 2015 IBM Corporation
 Integration Patterns for Virtual MDM Implementations

SPECIALIZED INTEGRATIONS

30 © 2015 IBM Corporation

Unstructured Text Correlations
Analyze Unstructured text sources and manage associations with master data

How It Works Correlate to

• Discover linkages between text and relevant
master data entities Investigate
• Link to additional information
• Investigative tool for text analysis Unstructured Text
• Complete, accurate and timely views

Client Value
• Enhanced customer service
• Quicker, more accurate issue resolution
• Increased upsell/cross opportunity

Usage
• Enhance detection of relationships between
entities
• Enhance entity resolution from the evidence
hidden within the documents.
• Enrich knowledge base by adding additional
information to MDM records

Using MDM in fraud investigations

31 © 2015 IBM Corporation
 Link unstructured text to master record and discover additional
relationships – Extensions for Unstructured Text Correlation

Example: Public Safety

4. Matched entities identified; no
automatic resolution of master record

1. Potential suspect identified

2. Search master records in

InfoSphere MDM 3. Uncover links between entities from
documents, emails, blogs etc.

32 © 2015 IBM Corporation

 On-premise Virtual MDM to Cloud Integration
Cast Iron Live InfoSphere MDM Supercharges
Orchestration APIs Salesforce.com serving LOB

MDM Services
• Enrich SFDC ‘Account’
TIP information by leveraging the
broader enterprise master
Data Refinery/Bluemix information in InfoSphere MDM

• Improved SFDC ‘Account’

search and remediation &
support for both virtual and
physical hubs

• Leverages WebSphere Cast Iron

Cloud Integration on Information
Server pack for SFDC

• MDM Services TIP ( Template

Integration Pattern ) in Cast Iron
tooling

Cloud Integration for Bluemix :

Provides the on-ramp for on-prem
MDM Services for Data Refinery
APIs

• MDM’s Cast Iron Live

Orchestrations APIs in Bluemix

• Gateway to Data Refinery for on

premise MDM APIs
* v1.5.0.2 needed

• Developer appeal
33 © 2015 IBM Corporation
 Integration Patterns for Virtual MDM Implementations

CLOSING NOTES

34 © 2015 IBM Corporation

Feedback and connect

 Provide Feedback to organizers on your preferences for deep dive topics in one or more
integration areas you are interesting in

 Contact me if you are dealing with issues or specific opportunities where you need help
 [email protected]

35 © 2015 IBM Corporation

  BPM component based
In Summary … Stewardship and
remediation
 Finance  Master Data Policy  Billion Scale
 Government Management Matching
 Retail  Extend Master Data  Social Data
 Telco Governance to Business Integration
 Insurance Applications  Enterprise Free
 Travel & Transportation  Support for Mobile Apps Text Search
 More …  Data Integration
on-ramps for Big
Data Analytics
 Virtual Line of Business  Cloud
 Physical oriented Integration
 Hybrid  IaaS
Industries  PaaS /
Big Data Ready DataWorks
 SaaS
Solutions
from BPs

Implementation
Cloud Enabled
Styles

Acquire, Manage, Deliver

Critical Information Assets

36 © 2015 IBM Corporation

37 © 2015 IBM Corporation
 Integration Patterns for Virtual MDM Implementations

BACKUP SLIDES

38 © 2015 IBM Corporation

For details on Clover ETL, review the following links

 InfoSphere MDM SE Documentation

http://www-01.ibm.com/support/knowledgecenter/SSWSR9_11.4.0/com.ibm.swg.im.mdmhs.mdmstages.doc/topics/r_migratingfromclovertodatastage.html

 If you are upgrading, please contact your local IBM rep and keep Todd Weise ( [email protected] ) on
copy

 We are continuing our partnerships with Javelin - Clover ETL’s external statements
http://www.cloveretl.com/ibm-mdm

http://www.cloveretl.com/content/cloveretl-guarantees-support-ibm-infosphere-mdm-users-and-enhances-their-data-processing-pos

39 © 2015 IBM Corporation

Redbook to know more about Integrating SaaS Applications and
MDM

http://www.redbooks.ibm.com/redbooks.nsf/RedpieceAbstracts/sg248220.html?Open
`

40 © 2015 IBM Corporation