2. 2 Business Intelligence Architecture
• Big Data / Analytics / BI & Cloud Solutions Specialist
• http://www.linkedin.com/in/JulioPhilippe
• Skills
WHO AM I
Big Data
Analytics
Business Intelligence
Data Warehousing
IT Transformation
IT Solutions
Cloud Computing Datacenter
Optimization
Business Development
Architecture
Hadoop
Management
Mentoring
3. 3 Business Intelligence Architecture
« The data are not created relevant,
they become though! »
4. 4 Business Intelligence Architecture
Manufacturer / Supplier collaboration
Reengineering Sales
Reengineering Distribution
Intensified focus on customer
Internationalization Reduce time to market
Manufacturing Telecommunications Information & Technology
Regulations observance
Increase risks management
Create Intelligent Buildings
Provide timely access to decision support
Accomplish more work with fewer
resources
Banking & Finance
Education & Research
Government
Collecting and Sharing experiences
Improved lifecycle management of products
Enhanced information dissemination
Focus on products that have the best chance of
getting to market
Pharmaceutical
Transportation & Travel
Retail
Need to manage profitability and control
expenses
Increase refining capacity in traditional
petroleum
Investments in the renewable energy sector
Energy
Media & Entertainment
Evolving consumer behaviors
Healthcare
Accelerating Employers-Led Initiatives
New Consumers-centric Technologies
SOME KEY BUSINESS DRIVERS
5. 5 Business Intelligence Architecture
• Backlog of report requests and need to provide self-
service access to intelligence
• Resource constraints inhibit ability to broaden BI
deployment to meet demand
• Complex IT environments – Disparate data,
applications, and integration challenges
• Push to reduce costs and standardize tools
KEY IT DRIVERS
6. 6 Business Intelligence Architecture
• Costs effective to purchase, support and manage and proven ROI
• Effortless to design and implement
• Leverage your existing investment
Costs / Complexity
• Scale without impacting the response time
• To handle the data volume and throughput demands of users
• Scalable infrastructure that supports growth
Scalability / Performance
• BI solutions able to provide information to users when they need it
• Extensive non-disruptive services and upgradeAvailability
Flexibility • Server, Storage and network integration
• Dynamic infrastructure with virtualization technologies
VALUE PROPOSITION
7. 7 Business Intelligence Architecture
DATA MODEL
• OLTP Model
- Transactional
- Third Normal Form
• OLAP Model
- Business Intelligence
- Star Schema
- Time Management
- Aggregation
Product
Id-product
Name
Sale
Id-product
Id-company
Price
Margin
Creation date
Customer
Id-customer
Name
Region
Id-region
Name
Tables
Sale
Id-product
Id-customer
Id-period
Id-region
CA
Margin
Extraction Transformation Load
Region
Id-region
Name
Period
Id-period
Name
Customer
Id-customer
Name
Fact TableDimension
Tables
Product
Id-product
Name
8. 8 Business Intelligence Architecture
• Fast Predictable Performance
- Monitor and ensure database performance with
automated diagnostics and tuning
• Lower Ongoing Costs
- Drive down the cost of operations with
automated change and configuration management
• The Fastest Time to Value & Lowest Risk
- Automate testing of patches, changes and upgrades
while keeping data secure
MANAGEMENT, ADMINISTRATION & SUPPORT
9. 9 Business Intelligence Architecture
NEW BUSINESS INTELLIGENCE
Storage Trends
New Data Structure
Distributed File Systems, NoSQL Database, NewSQL…)
Compute Trends
New Analytics
(Massively Parallel Processing, Algorithms…)
Proprietary and dedicated
data warehouse
OLTP is the
data warehouse
General purpose
data warehouse
Objects storage
Distributed File Systems Federated/
Sharded
Master/Master
Master/Slave
Enterprise
data warehouse
Multi-structured
Data
Logical
data warehouse
Master Data Management, Data Quality, Data Integration
10. 10 Business Intelligence Architecture
MOVING COMPUTATION TO STORAGE
Moving Data processing to Storage
Application
Data Processing
Storage
Legacy
Application
Data Processing
Metadata Mgmt
Storage
Emerging
Application
Data Processing
Metadata Mgmt
Storage
Next Gen.
Metadata Mgmt
Storage Array (SAN, NAS) Servers
Network
General Purpose Storage Servers
• Combine server with disks & networking for reducing latency
• Specialized software enables general purpose systems designs to
provide high performance data services
11. 11 Business Intelligence Architecture
DATA WAREHOUSE
• Data Warehouse Appliances
– EMC Greenplum
– Microsoft Parallel Data
Warehouse
– IBM Netezza
– Oracle Exadata
– SAP HANA
– ParAccel Analytic Database
– Teradata
– HP Vertica
– …
• SQL Database
• Massively Parallel Processing
• Hadoop Connectivity
• Column-Oriented database
• In-Memory database
12. 12 Business Intelligence Architecture
Extract volume
Complexity
Data Volume
Concurrent Users
Concurrent Requests
Data Processing
Complexity
Batch
Time Range
S/W Components
DB Structure
CostsUsers
Time Range Benchmarks
Results
CPUs Memory Disks
A balanced configuration from Processors to Disks
I/O
BI SIZING METHODOLOGY
13. 13 Business Intelligence Architecture
• What is the Web servers sizing ?
– http server, proxy server, application server, portal server, LDAP server
• What is the ETL server sizing ?
• What is the Database servers sizing ?
– Staging area, data warehouse, data marts
• What is the backup server sizing ?
• What are the key elements sizing ?
– User, data, time range, SLA…
KEY QUESTIONS
14. 14 Business Intelligence Architecture
DWH
Operational Data Source
DATA MARTS
%index, %metadata
%index, %aggregates, %axis
%Index
Users time range
h1 – h2
ETL Flow
ETL Flow
* Usable data, indexes, aggregates
and metadata included...
n1 Low users n2 Medium users n3 High users
ETL time range
h1 - h2
WORK
TEMP, LOG...
External Flow
Data Processing
complexity
Data Processing
Complexity
SA
% Aggregats
re-building
DATA FLOWS DIAGRAM
15. 15 Business Intelligence Architecture
• Software (e.g. Oracle DB , Sybase DB, DB2, MS SQL Server…)
• Raw data are the data sources resulting from the operational
systems (e.g. CRM, RH, Billing, Purchases, SCM...)
• Usable data are raw data, indexes, aggregates, metadata,
dimensions, indicators and data work
• The database is structured on 3 levels
- Staging Area for data validation
- Data Warehouse for data reference
- Data Marts (cubes) for data analytics
DATA VOLUME QUALIFICATION
16. 16 Business Intelligence Architecture
• Software (e.g. SAP BO Explorer, Oracle BI, SAS analytics…)
• Number of users have access to BI system (named users)
• Number of users have access to BI system simultaneously
(concurrent users)
• Users are classified by typology
- Low users view predefined and static reports
- Medium users navigate within reports, do slicing and dicing, but
usually hit aggregates
- High users run ad-hoc queries with a high probability of full table
scans
• Operational period (e.g. from Monday to Friday)
• Time range (e.g. 08:00 AM - 07:00 PM)
USERS QUALIFICATION
17. 17 Business Intelligence Architecture
• Software (e.g. Oracle Data Integrator, PowerCenter
Informatica…)
• ETL processing are classified by typology
- Simple processing represents simple calculations and
concatenations
- Medium processing represents average calculations, medium
concatenations
- High processing represents heavy calculations, statistical, complex
algorithms and heavy concatenations
• Data retention duration (e.g. year, month, day)
• Operational period (e.g. from Monday to Friday)
• Operation frequency (e.g. daily, weekly, monthly)
• Time range (e.g. 08:00 PM to 06:00 AM)
DATA INTEGRATION QUALIFICATION
18. 18 Business Intelligence Architecture
BI ARCHITECTURE
Primary Site Secondary Site
SAN Switches
Storage
Array Data Replication
DB Servers
LAN Switches
Apps Servers
Storage
Array
• SQL based
• High availability
• Enterprise database
• Right design for structured data
• Hardware storage (SAN, NAS, DAS)
19. 19 Business Intelligence Architecture
NEW BI ARCHITECTURE
Network Switches
Data Nodes – Compute and storage integrated
• Not only SQL based
• High scalability, availability and flexibility
• Compute and storage in the same box for reducing the network latency
• Right design for semi-structured and unstructured data
Apps Servers