Contenu connexe
Similaire à 6 enriching your data warehouse with big data and hadoop (20)
Plus de Dr. Wilfred Lin (Ph.D.) (20)
6 enriching your data warehouse with big data and hadoop
- 2. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.2
Agenda
Opportunity
Challenges
Strategy
Examples
Best Practices
- 3. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.3
Connecting With Your Customer
- 4. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.4
Big Data Improves Operational Performance
Source: Economist Intelligence Unit, .”The Deciding Factor: Big Data and Decision Making“
Big data benefits
seen growing
substantially
- 5. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.5
Use
Data
12%
Executives who feel they
understand the impact data
will have on their organizations
Produce
Data
The Problem
- 7. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.7
DON’T BELIEVE
EVERYTHING
YOU READ
- 8. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.8
Electricity: AC or DC?
- 9. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.9
RDBMSRDBMS
Today
Discovery & Analytics
Business Intelligence
- 10. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.10
RDBMSRDBMS
Today
Discovery & Analytics
Business Intelligence
External
ETL Cluster
- 11. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.11
RDBMSRDBMS
What about Archived Data?
Discovery & Analytics
Business Intelligence
Archive
External
ETL Cluster
- 12. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.12
RDBMSRDBMS
What about New Data?
Discovery & Analytics
Business Intelligence
?
Archive
External
ETL Cluster
- 13. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.13
RDBMSRDBMS
Expand Your Data Warehouse
Discovery & Analytics
Business Intelligence
?
External
ETL Cluster
Archive
- 14. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.14
RDBMS
Hadoop
RDBMS
Hadoop
Integrate Hadoop and RDBMS
Discovery & Analytics
Business Intelligence
- 15. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.15
“[Facebook] started in the Hadoop world. We are now
bringing in relational to enhance that. We're kind of going [in]
the other direction ... We've been there, and [we] realized
that using the wrong technology for certain kinds of
problems can be difficult.”
Ken Rubin
Director of Analytics
Facebook
http://tdwi.org/Articles/2013/05/06/Facebooks-Relational-Platform.aspx?Page=1
- 16. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.16
Get Fast Answers
to New Questions
Create a Data
Reservoir
Predict More,
More Accurately
Accelerate
Data-Driven Action
Key Hadoop Use Cases
Complementing An Existing Data Warehouse
ETL
- 17. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.17
• Transactional data
• Customer information
• Web log and session data
• Machine/Sensor data
• Historical data
Data Reservoir
Keep All Potentially Valuable Data in One Place
- 18. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.18
RDBMSRDBMS
Today
Discovery & Analytics
Business Intelligence
?
Archive
External
ETL Cluster
- 19. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.19
RDBMS
Hadoop
RDBMS
Hadoop
Create an Active Archive with Hadoop
Discovery & Analytics
Business Intelligence
Σ
- 20. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.20
80 years of historical data in Hadoop
Structured and unstructured data includes
customer data, economic trends, telematic
sensors, weather, public data
Integrated with mainframes and EDWs
Before Hadoop, could analyze only one
state, took 24 hours
With Cloudera, can analyze risk across all
50 states, in 16 hours (500x improvement)
First 3 use cases: Data hub, ETL offload,
advanced analytics
Comprehensive risk analysis
Customer Example: Insurer Real-Time Data Hub
Cloudera
Hadoop
EDW and
Mainframe
Customer
Data
- 21. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.21
Optimize offers
Increase revenue
Lower costs (offer bytes?)
Reduce complexity
Faster time to value
Maximize offer effectiveness
Customer Example: Travel Industry
Big Data
Appliance
Legacy
Data
Warehouse
Customer
Data
- 22. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.22
Fast access to 85% more data
Lowered costs
Simplified architecture
Faster time to value
Compliance, cost reduction
Customer Example: Regional Bank
Big Data
Appliance
Oracle
Exadata
Mainframe,
RDBMS
Oracle Data Integrator
- 23. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.23
• Batch window constraints
• Adding value vs. adding cost
• Analysis vs. Transformation
• Analysis vs. Data movement and replication
• Uncertain value of new data sources
ETL Challenges Today
- 24. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.24
Discovery & Analytics
Business Intelligence
RDBMSRDBMS
Typical ETL Today
External
ETL Cluster
- 25. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.25
RDBMS
Hadoop
RDBMS
Hadoop
ETL Offload with Hadoop
Discovery & Analytics
Business Intelligence
Σ
- 26. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.26
Exponential growth in data,
generated by new consumer devices
ETL and storage constraints limited
analytics to 1% sample
Now combined Oracle Exadata and
Cloudera Hadoop delivers analytics
on 100% of data (half a PB per day!)
Query times reduced dramatically
(i.e. from 4 days to 53 minutes)
90% reduction of ETL code base
From 1% sampling to 100% analysis
Customer Example: Communications Services
Archive Storage
Data Warehouse
Complex
Correlation
Alerting
Filter
&
Split
Event
Monitoring
Streaming ETL
Streaming ETL
Teleco
m
Services
Before
- 27. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.27
Exponential growth in data,
generated by new consumer devices
ETL and storage constraints limited
analytics to 1% sample
Now combined Oracle Exadata and
Cloudera Hadoop delivers analytics
on 100% of data (half a PB per day!)
Query times reduced dramatically
(i.e. from 4 days to 53 minutes)
90% reduction of ETL code base
From 1% sampling to 100% analysis
Customer Example: Communications Services
Archive Storage
Data Warehouse
Complex
Correlation
Alerting
Filter
&
Split
Event
Monitoring
Streaming ETL
Streaming ETL
Teleco
m
Services
Before
Data
Warehouse
Alerting
Filter
&
Split
Event
Monitoring
Hadoop
Archive Storage
ETL
Correlation
Stage 1 DWH
Teleco
m
Services
After
- 28. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.28
Identifying Potential Opportunities
Do you have a
problem with ETL
performance?
Do you have
potentially valuable
data that you aren’t
using, but might
deliver new insight?
Should you focus
on analyzing
structured data,
unstructured data,
or a combination?
Are Big Data
solutions already
being built as silos?
- 29. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.29
Challenges in the Way
& Oracle Strategies
CHALLENGES ORACLE STRATEGY
• Fragmented Solutions • Specialized but integrated data stores and tools
• Difficulty of Self-Service BI • Flexible, guided, automated BI & data discovery
• Data Not Current • Solutions for Just-in-Time well defined data
• Time to ROI / Development Time • Horizontal & industry pre-built solutions, engineered systems
• Growing Diversity of Data & Users • Enterprise solutions for 1000s diverse users, petabytes data
• Manageability, Security, Cost • Centrally managed with advanced security / governance
- 30. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.30
Big Data Readiness
The R&D Prototype Stage
Skills needed
– Distributed data deployment (e.g. Hadoop)
– Python or Java programming with MapReduce
– Statistical analysis (e.g. R)
– Data integration
– Ability to formulate business hypotheses
– Ability to convey business value of Big Data
- 31. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.31
Are You Ready for Big Data?
Contact Your Account Team
Have you delivered mature analytics solutions for
structured data?
Can Big Data make a difference to the business?
Have you built a Big Data prototype, built skills, and
proved value?
Do you have an enterprise integration & deployment
strategy for Big Data?
- 32. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.32 Copyright © 2013, Oracle and/or its affiliates. All rights reserved.32
Continuous
Innovation
Big Data at Work
- 35. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.35
RDBMS
Hadoop
RDBMS
Hadoop
Master diagram
Discovery & Analytics
Business IntelligenceExternal
ETL Cluster
Data
Mart
Data
Mart
Archive
ΣΣEvent
Processing
- 36. Copyright © 2013, Oracle and/or its affiliates. All rights reserved.36
Discovery & Analytics
Business Intelligence
RDBMSRDBMS
Master diagram without Hadoop
External
ETL Cluster
Data
Mart
Data
Mart
Archive
ΣEvent
Processing