ON-DEMAND VISIBILITY
Data Virtualization an Overview
Ian Pestell
Director EMEA Operations
Copyright © 2005 Composite Software, Inc. All Rights Reserved.
Is Your Data an Asset or a Liability?
“Managers spend two hours a day looking for
information they need, and almost half the
data is useless once they get it”
Source: Accenture
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
2
Data within SOA
SOA
SOA Architectures implement
Business Process
Through Application Logic
Operating on Data
Business
Business
Process
Process
Application
Application
Logic
Logic
Data from many sources
Existing Legacy systems
Databases, Web Services,
Files
Complicated by package
applications, ERP and CRM
Data
Data Services are often
addressed late in SOA
Legacy
Data
Application
Data
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
3
Data Integration Options
Business
Solutions
Constant
Change
Dashboards
Data
Integration
Challenge
Reporting
SOA
Code within the Application?
File Extracts / Database Replications?
ETL / Physical Warehouses, Marts, Stores?
EII / Virtual?
Existing
Data
Packaged
Apps
RDBMS
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
Data
Warehouses
4
Web
Services
Flat Files
Siloed
&
Complex
Data Consolidation / ETL
Physical data movement via
scripts
Application
Data is extracted from source
systems
Transformed
Data Cleansing, OLTP to OLAP
Data
Warehouse
Loaded into target system
Physical Data Warehouse,
Mart, Operational Data Store
Very large volumes of data
High degree of transformation
Data is only as current as last
load
MDM
Hub
Data Mart
/ ODS
ETL
Server
Files
Databases
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
5
Applications
Data Federation / EII
Virtual data views via modelling
Data is left in-situ within data
sources
EII server federates and
transforms data on demand
Application
No Data Warehouse required
High degree of flexibility in adding
new data sources
Lower development & operating
costs
cache
EII
Server
Databases
Applications
Moderate to high data volumes
Low to medium transformations
Data is ‘near’ real-time
Files
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
6
Solutions to Data Integration – Application
Integration
Application
Logic
Integration performed in the
application logic
Data sources require WS API
Vendor – bespoke
Access to WS API via ESB
Enterprise Service Bus
Supported by standards
ws-orchestration, ws-choreography
Disadvantages
WS API
Legacy
Data
WS API
WS API
WS API
Application
Data
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
7
Complex logic with Application
High develop time
Maintenance of change
Enabling WS APIs for data
High load on ESB
Latency of ESB
Solutions to Data Integration – Data Services
Layer
Application
Logic
Implement Data Services
Layer
Provides set of reusable
data services
WS access into ESB
Connects to data sources
using native interfaces
Integration / optimization
only done within data layer
Reduces load
Enterprise Service Bus
Data Services Layer
Legacy
Data
Application
Data
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
8
You Can Integrate Data Differently
Business
Solutions
Real-time
Information
Dashboards
Reporting
Applications
Positions
Virtual
Data
Layer
Orders
POs
Virtualizes
Trades
Vendors
Abstracts
Invoices
Customers
GLs
Employees
Federates
Existing
Data
Packaged
Apps
RDBMS
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
Data
Warehouses
9
Web
Services
Flat Files
Siloed
&
Complex
Composite’s Mission
Business
Applications
Dashboards
Reporting
SOA
Composite helps IT teams who need to build new applications faster,
by providing data services middleware that lets them access any
data in the enterprise… and deliver it on demand.
Existing
Data
Packaged
Apps
RDBMS
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
Data
Warehouses
10
XML Docs
Flat Files
What We Do
Operational Reporting
Access
Combine
Deliver
>>> On Demand <<<
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
Single Source for all Data
Standardized Access
Reusable Data Services
Data Services
(SOA)
NA Sales
Data Engine
For
Operational BI
Drug Discovery
Order Management
Portfolio Management
Inventory Management
Information
Infrastructure
App
11
Composite Software at a Glance
Privately-held, Silicon Valley firm
Strong Business Model & Team
Best-of-Breed Virtual Data Integration / EII
Discover, Virtualize, Abstract & Federate
IE 2008
Editor’s Award
Proven Business and IT Value
Increased Flexibility
Greater Agility
Lower Costs
Financial Services
Top
5
Inv
Ban
k
Top 5
Comm
Bank
Top 5
I-Bank
Top 5
I-Bank
Top 5
I-Bank
Mutual
Fund
Life Sciences
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
Technology
Energy and Utility
Consumer/Retail
Life Sciences
Public Sector
12
How we do it – Data Views
Composite View
Step 1 - Development:
Build and Publish a View
(A View contains Metadata
on the actual data source –
format, type, location,
access, etc.)
Data Source 1
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
Data Source 2
13
How we do it – Data Views
Reporting
Analytics
Step 3 – Runtime:
Deliver Information
(Combined data creates
meaningful information)
Composite
Applications
Composite View
Step 1 - Development:
Build and Publish a View
(A View contains Metadata
on the actual data source –
format, type, location,
access, etc.)
1
A
B
a
b
2
C
D
c
d
3
E
F
e
f
4
G
H
g
h
Step 2 – Runtime:
Access and Combine
(Queries are optimized
for speed and efficiency)
Data Source 1
Data Source 2
1
A
B
1
a
b
2
C
D
2
c
d
3
E
F
3
e
f
4
G
H
4
g
h
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
14
Composite Lets You Easily Construct Data Services
Access Data
Request Results
Combine Data/
Model
Access
Data
Auto-Generated
Deliver
Results
WSDL
Combine
Data/
Model
Request Results
Deliver Results
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
Auto-Generated WSDL
15
Composite Information Server Architecture
Easily Works Within Complex Environments
Client and Enterprise Applications
Web Services (SOAP/JMS)
Relational SQL (ODBC/JDBC/ADO.NET)
Composite Studio
COMPOSITE INFORMATION SERVER
Modeler
Composite Web
Manager
Administration API
Manager
Security
Data Services
Pre-Built
•S.
Customer-Built
(Views, SQL Script, XQuery, Java)
Cache
Query Processing Engine
Security
Relational (SQL) WS (SOAP/JMS)
RDBMS
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
Excel
Java
APIs
MF Adapter
Web Service Flat File Custom Application
16
Mainframe
Why Composite Application Data Services™
Excel
BI Tools
Composite
Apps
SQL / WS
Complex API
Adaptor
Application
Views
Enterprise Data
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
17
Data Services are a Critical SOA Component
Composite Applications
1) Data services stand alone
2) Transaction services stand
alone
3) Data services integrated
with transaction services
4) Data services with
orchestration using a
business process
management tool or ESB
5) Data services WSDL stored
in a UDDI registry
Orchestration
ESB / BPM
Registry
Services
DataData
Services
(Info Server)
“Over 60% of SOA
services will be
data services”
Transaction Services
Info Server
App Server
Source: IDC
Packaged Apps
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
SOAP/HTTP
XML Files
18
XML/HTTP
Files
Custom
Apps
NY Investment Bank – New e-business
application
Implementation of new ebusiness application
Initial estimate 2 man
years
1 year application
1 data integration
Using EAI approach
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
19
NY Investment Bank – New e-business
application
Implementation of new e-business
application
Initial estimate 2 man years
1 year application
1 data integration
Using EAI approach
Using Composite Data services
Reduces Data integration from 12
Month to 4 Months
Created optimized and re-usable
data services
Introduced abstraction layer
between application and data
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
20
Data Services At Work - New York Bank
Multiple Projects
1.
2.
3.
4.
Business and IT Benefits
Prime Brokerage Reporting
Risk Management
Back Office Reconciliation
Portfolio Balancing
Data Required (100%)
Reuse (nearly 40%)
Lower Impact (75% or more)
Faster Response (up to 80%)
Data
Services
Layer
Existing
Data
Packaged
Apps
RDBMS
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
Data
Warehouses
21
XML Docs
Flat Files
Single View of Payroll and Variances
Husasmidjan, Inc.
Situation & Requirements:
Scandinavian construction materials
retailer/wholesaler, 25 stores
Business requirement: detect and address
variances between planned and actual payroll
expense
Business user: Office of the CFO
Technical requirement: Expose and combine
SAP R/3 and in-house budget data
Composite
Information
Server
Solution:
Rollup View
1. Create views to SAP payroll information
2. Create views to other payroll information
3. Build roll-up view for analysis
Local views normalize
across different stores
and apps
4. Expose view into budget system
Benefits:
More accurate variance information sooner
Rapid application development
Payroll data sources
Legacy
Files
Copyright © 2006 Composite Software, Inc. All Rights Reserved.
22