Appliancebased Approaches to Data Warehousing and Business Intelligence

1
Appliance-based Approaches to Data Warehousing
and Business Intelligence
January 14, 2009
Don Dybas Sales Consulting Manager Oracle Public
Sector
2
Getting on the Same Page IT Lines of Business
Competing Agendas
3
The Evolving Role of BI
To
From
4
1. Pervasive Use NYS Medicaid DW
Financial Analysis
Program Eligibility
Planning Apps
Federal Reporting
Data Warehouse
Partners
Web Servers
Consumers
Internal End Users
Financial Performance
Clinical Analysis
Program Design
Fraud Detection
5
2. Real-Time Predictive Infrastructure
Data Marts Reporting
Data Warehouse
Applications
Eligibility
Delivery
Self Service
DW
Call Center
Business Intelligence
Financials
Providers
Eligibility
Claims
ETL
6
3. A Unified, Enterprise View
Users BI tool of choice
Microsoft Excel
Hybrid schema Relational OLAP databases unite
In-place data mining
Open standards-based platforms
Database
7
4. Insight Driven Business Processes
8
5. Pre-Built Analytic Solutions Value
BI Applications (Payroll, Personnel, Financial,
etc.)
Build from Scratchwith Traditional BI Tools
Training Roll-out
Define Metrics Dashboards

• Faster deployment
• Lower TCO
• Assured business value

DW Design
Back-end ETL andMapping
Role-based dashboards and alerts Thousands of
pre-defined metrics
Training Rollout
Specific Metrics Dashboards
Prebuilt DW design Adaptable to enterprise DW
DW Design Mod
Pre-built ETL Tailoring
Prebuilt ETL to source systems
Weeks or months
Quarters or Years
9
6. Enterprise Performance Management
Visualize
Analyze
Publish
Integrate Data
Model
Collaborate
Monitor
Secure Manage
Compare
Predict
Sense
Respond
10
Some Implications for IT

• A. More data is on-line
• B. Data growth is accelerating
• C. Balanced platforms are needed to support
  growth

11
A. More On-Line DataInformation Lifecycle
Management
12
B. Explosive Data GrowthReal Projected Data
Volumes at Amazon.com
- Additional lines of business / product lines
supported - Standard reporting growth, more
partners supported
13
C. Balanced Platform ConfigurationsAppliances
Appliance Foundations

• You determine the potential size workload type
• You determine initial sizing choices
• Platform vendors determine final configuration

14
But what about these folks?Business Drivers
could Change
15
Extreme Data WarehousingAddressing High Data
Volume and Pervasive Use
Custom
Optimized Warehouse
Reference Configurations
Data Warehouse Appliances

• Complete Flexibility
• Any OS, any platform
• Easy fit into a companys IT standards
• Multiple HW partners

• Documented best-practice configurations for data
  warehousing
• Multiple HW partners

• Scalable systems pre-installed and
  pre-configured ready to run out-of-the-box
• Multiple HW partners

• Highest performance
• Pre-installed and pre-configured
• Sold by Oracle and Others

16
What is a Data Warehouse AppliancePre-Configured
High Performance Machine

• Several Database servers
• Multi-core Chipsets, 4-1 RAM-Core Ratio
• Operating System (e.g. Linux)
• Grid-enabled RDBMS (e.g. Oracle)
• Several Storage Servers (SAS or SATA)
• Hundreds of TB raw storage
• High-Speed InterConnect Storage switches
• LAN Switch
• Keyboard, Video, Mouse (KVM) hardware
• Hardware Install at Customer
• Software Hardware
• Pre-Installed Pre-Configured
• Hardware Warranty Support
• Parts / Labor / On-site
• 24X7, 4 Hour response time

Add more racks for unlimited scalability
17
Extreme Data WarehousingAddressing High Data
Volume and Pervasive Use

• Data warehouse appliances address data bandwidth
  issues
• Lots of I/o paths,
• Balanced I/o, CPU, memory
• Grid / HA Configurations
• However, Some Warehouse Appliances
• Are non-standard
• Require specialized staff
• Use proprietary hardware
• Cant handle mixed OLTP and Reporting

18
Extreme Data WarehousingAddressing High Data
Volume and Pervasive Use

• A successful warehouse appliance must
• Scale to meet business requirements
• Long-term architecture
• Minimize re-platforming
• Fit existing s/w and h/w technology skillsets
• Standards reduce operating costs
• Handle mixed workloads
• High-Volume Batch / Scanning
• Traditional Reporting
• Affordable and efficient

19
Extreme Data WarehousingAddressing High Data
Volume and Pervasive Use

• A successful warehouse appliance must
• Complement and support ubiquitous BI client
  platforms
• Some dw load patterns can be anticipated
• Some are un-predictable and governance-challenged
• Drive i/o efficiencies to the extreme
• Smart scanning
• Function-shipping
• Storage-based data filtering
• Support High Availability
• Server Clustering
• On-line Operations
• Partitioned Data Sets
• Complementary Storage-based mirroring

20
Extreme Data WarehousingAddressing High Data
Volume and Pervasive Use

• A successful warehouse appliance must
• Support DW / DM Consolidation
• Address Data Governance Separation of Duties
• Address Mixed User Workloads and Classes
• Support non-traditional data storage and access
  patterns
• Geo-Spatial Geo-Raster
• Structured, Unstructured, and Semi-Structured XML
• Documents, Images, and other Binary Objects
• Green Storage
• Information Lifecycle Management
• Advanced Compression / De-duplication
• Enterprise Search

21
Questions