Creating Interactive OLAP

Applications with MySQL

Enterprise and Mondrian

Julian Hyde: Chief Architect, OLAP, at Pentaho

and Mondrian Project Founder

Wednesday, April 16th 2008

Agenda

Pentaho Introduction
Mondrian features and architecture
Schemas and queries
olap4j
Roadmap
Case Studies
Business Intelligence suite
Q&A
Pentaho Introduction
World’s most popular enterprise open source BI Suite
2 million lifetime downloads, averaging 100K / month
Founded in 2004: Pioneer in professional open source BI

Management - proven BI and open source veterans

from Business Objects, Cognos, Hyperion, JBoss, Oracle, Red Hat, SAS

Board of Directors – deep expertise and proven success in open source

Larry Augustin - founder, VA Software, helped coin the phrase “open source”
New Enterprise Associates – investors in SugarCRM, Xensource, others
Index Ventures – investors in MySQL, Zend, others

Widely recognized as the leader in open source BI

Distributed worldwide by Red Hat via the Red Hat Exchange
Embedded in next release of OpenOffice (40 million users worldwide)
What is OLAP?

View data “dimensionally”

i.e. Sales by region, by
channel, by time period

Navigate and explore

Ad Hoc analysis
“Drill-down” from year to
quarter
Pivot
Select specific members for
analysis

Interact with high

performance
Technology optimized for rapid
interactive response
Mondrian features and architecture
Key Features

On-Line Analytical Processing (OLAP)

cubes
automated aggregation
speed-of-thought response times
Open Architecture
100% Java
J2EE
Supports any JDBC data source
MDX and XML/A
Analysis Viewers
Enables ad-hoc, interactive data
exploration
Ability to slice-and-dice, drill-down,
and pivot
Provides insights into problems or
successes
How Mondrian Extends MySQL for OLAP
Applications
MySQL Provides Mondrian Provides

Data storage Dimensional view of data

SQL query execution MDX parsing
Heavy-duty sorting, correlation, SQL generation
aggregation Caching
Integration point for all BI tools Higher-level calculations
Aggregate awareness
 Microsoft Excel
(via Spreadsheet
Open Architecture Viewers
Services)

Open Standards (Java, XML,

MDX, XML/A, SQL) Web Server
JPivot servlet
Cross Platform (Windows &
Unix/Linux)
J2EE Application Server

J2EE Architecture JPivot servlet XML/A servlet

Server Clustering Mondrian

Cube Cube Cube

Fault Tolerance Schema Schema Schema

XML XML XML cube cube cube

File or RDBMS Repository

Data Sources JDBC JDBC JDBC

JDBC
JNDI

RDBMS RDBMS
<mondrian/jpivot demonstration>
Schemas and queries
A Mondrian schema consists of…

A dimensional model (logical)

Cubes & virtual cubes
Shared & private dimensions
Calculated measures in cube and
in query language
Parent-child hierarchies

… mapped onto a star/snowflake

schema (physical)
Fact table
Dimension tables
Joined by foreign key relationships
 Writing a Mondrian Schema

Regular cubes, dimensions, <!-- Shared dimensions -->

<Dimension name="Region">
hierarchies <Hierarchy hasAll="true“
allMemberName="All Regions">
Shared dimensions <Table name="QUADRANT_ACTUALS"/>
<Level name="Region" column="REGION“
Virtual cubes uniqueMembers="true"/>
</Hierarchy>
Parent-child hierarchies </Dimension>
<Dimension name="Department">
Custom readers <Hierarchy hasAll="true“
allMemberName="All Departments">
Access-control <Table name="QUADRANT_ACTUALS"/>
<Level name="Department“
column="DEPARTMENT“
uniqueMembers="true"/>
</Hierarchy>
</Dimension>

Refer to http://mondrian.pentaho.org/documentation/schema.php )
Tools

Schema
Workbench
Pentaho cube
designer
cmdrunner
 MDX – Multi-Dimensional Expressions

A language for multidimensional queries

Plays the same role in Mondrian’s API as SQL does in JDBC
SQL-like syntax

SELECT {[Measures].[Unit Sales]} ON COLUMNS,

{[Store].[USA], [Store].[USA].[CA]} ON ROWS
FROM [Sales]
WHERE [Time].[1997].[Q1]

… but un-SQL-like semantics

(Refer to http://mondrian.pentaho.org/documentation/mdx.php )
olap4j
OLAP APIs

OLAP APIs that failed:

OLAP Council’s MDAPI and OLAPI
Sun’s JOLAP

OLAP APIs that succeeded:

Microsoft’s OLE DB for OLAP, ADOMD, XMLA

Mondrian has an API for creating running MDX queries:

Powerful and intuitive
Features the MDX language
Used by Mondrian’s XMLA provider, JPivot, other clients
But it’s Mondrian-only
The problem with APIs

Client 1 Client 2 Client 3

API 1 API 2 API 3

Server 1 Server 2 Server 3

The problem with APIs #2

Client 1 Client 2 Client 3

API 1 API 2 API 3

Server 1 Server 2 Server 3

The problem with APIs – the solution

Client 1 Client 2 Client 3

Common API

Driver 1 Driver 2 Driver 3

Server 1 Server 2 Server 3

olap4j

olap4j aims to be the “JDBC for OLAP”

An extension to JDBC
Also inspired by ADOMD.NET
Implementations for many OLAP servers
Enable one client to work against many servers
Break the ‘lock-in’
Encourage more businesses to ‘take a chance’ on open-source

Backed by:
Companies: Jedox, JasperSoft, Loyalty Matrix, LucidEra, Pentaho,
Tensegrity, Tonbeller AG
Projects: Halogen, JPivot, JRubik, Mondrian, OpenI, PALO
Community at SourceForge.net
olap4j connecting to mondrian in-process

import org.olap4j.*;

Class.forName("mondrian.olap4j.Driver");
OlapConnection connection =
DriverManager.createConnection(
"jdbc:mondrian:Jdbc=jdbc:mysql://localhost/foodmart;" +
"JdbcUser=foodmart;JdbcPassword=foodmart;" +
"Catalog=/WEB-INF/queries/FoodMart.xml;" +
"Role='California manager'");
OlapConnection olapConnection =
connection.unwrap(OlapConnection.class);
OlapStatement statement =
olapConnection.createOlapStatement();

OlapResult result =
statement.execute(
"SELECT {[Measures].[Unit Sales]} ON COLUMNS,\n" +
" {[Product].Members} ON ROWS\n" +
"FROM [Sales]");
olap4j connecting to SQL Server Analysis
Services via XMLA
import org.olap4j.*;

Class.forName("org.olap4j.driver.xmla.XmlaOlap4jDriver");
OlapConnection connection =
DriverManager.createConnection(
"jdbc:xmla:Server=http://marmalade/xmla/msxisapi.dll;" +
"Catalog=FoodMart;" +
"Role='California manager'");

OlapConnection olapConnection =
connection.unwrap(OlapConnection.class);
OlapStatement statement =
olapConnection.createOlapStatement();

OlapResult result =
statement.execute(
"SELECT {[Measures].[Unit Sales]} ON COLUMNS,\n" +
" {[Product].Members} ON ROWS\n" +
"FROM [Sales]");
 Web client #2
olap4j
Web client #1
Pure HTML AJAX client
web in web

architecture browser browser

Client:
Application server
In-process
Mondrian
In-process AJAX
XML/A JPivot
XMLA servlet
client server

HTML
olap4j
AJAX

olap4j driver
for Mondrian
Server: olap4j driver
for XML/A
Other olap4j
driver
Mondrian
(in-process)
Mondrian in-process
Mondrian via XMLA
Microsoft SSAS via
XMLA
Other OLAP server via
Mondrian
XMLA XML/A servlet
Microsoft SQL
Server Other XML/A
Analysis Server
Other OLAP server via Mondrian
Services

dedicated olap4j Data server #1 Data server #2 Data server #3

driver
olap4 components
Metadata
Driver
Driver Cube Level Property

management Dimension Member/

Measure
Set

Session Hierarchy Role Schema

Metadata
Session
MDX Connection MDX
MDX parser
Transform & Query
Transform
MDX object
layout model
Sort/Rank
Result set

Slice/Dice

Layout
Expand/
Collapse
Grid layout Chart layout
olap4j/mondrian roadmap

2006 September olap4j 0.5

(draft) 2007 August – mondrian 2.4

2007 October – olap4j 0.9 (beta)

2008 February – olap4j driver for 2008 March – mondrian 3.0

XMLA featuring olap4j API

2008 July – olap4j 1.0 2008 August – mondrian 3.1

featuring olap4j 1.0 API
mondrian 3.0 features

olap4j API
Rollup policy
Aggregate roles
Allow distinct-count measures to be aggregated
Improve dimension sharing
JNDI in connect string
Over 90 new MDX functions
Halogen

Pentaho incubator project

Slice/dice OLAP client
Built using GWT  AJAX
Built on olap4j  portable across OLAP servers
<halogen demonstration>
Case studies
Case Study: Frontier Airlines
Frontier Airlines
Key Challenges

“The competition is intense in the Understanding and optimizing fares to

ensure
airline industry and Frontier is
Maximum occupancy (no empty seats)
committed to staying ahead of the Maximum profitability (revenue per seat)
curve by leveraging technology that Pentaho Solution
will help us offer the best prices and Pentaho Analysis (Mondrian)
the best flight experience…. [the Chose Open Source RDBMS and
application] fits right in with our Mondrian over Oracle
philosophy of providing world-class 500 GB of data, 6 server cluster
performance at a low price.” Results
Comprehensive, integrated analysis to
set strategic pricing
Improved per-seat profitability (amount
not disclosed)
Why Pentaho
Rich analytical and MDX functionality
Cost of ownership
Pentaho at Loma Linda University Health
Care Leading Healthcare Provider
Key Challenges

“Pentaho Customer Support has been Providing analytics for billing and
operations supporting 500,000 patients
exceptional. This is a strategic and 600 doctors
application at LLUHC, and working Pentaho Solution
with Pentaho has accelerated our Pentaho Analysis Subscription
deployment and improved our overall Selected over Business Objects and
application delivery.” Cognos
Microsoft Windows Server with SQL Server
Integrated with LDAP and Active Directory
Results
Comprehensive analysis of time periods,
services provided, billing groups,
physicians
Centralized, secured, consistent
information delivery (versus prior Excel-
based system)
Ability to drill and analyze down to the
individual patient level
Why Pentaho
Open standards support and ease of
The big picture
Business Intelligence Suite

Mondrian OLAP
Analysis tools:
Pivot table
Charting
Dashboards

ETL (extract/transform/load)
Integration with operational reporting
Integration with data mining
Actions on operational data
Design/tuning tools
Pentaho Open Source BI Offerings

All available in a Free Open Source license

A Sample of Joint MySQL-Pentaho Users

“Pentaho provided a robust, open “We selected Pentaho for its ease-of-
source platform for our sales reporting use. Pentaho addressed many of our
application, and the ongoing support requirements -- from reporting and
we needed. The experts at OpenBI analysis to dashboards, OLAP and ETL,
provided outstanding services and and offered our business users the
training, and allowed us to deploy and Excel-based access that they wanted.”
start generating results very quickly.”
Next Steps and Resources

Contact Information
Julian Hyde, Chief Architect, [email protected]

More information http://www.pentaho.org and

http://mondrian.pentaho.org
Pentaho Community Forum http://community.pentaho.org
Go to Developer Zone
Discussions

Pentaho BI Platform including Mondrian

http://www.pentaho.org/download/latest
Mondrian OLAP Library only
http://sourceforge.net/project/showfiles.php?group_id=35302
Thank you for attending!