LIBeLIS - Java Forum Stuttgart

Download Report

Transcript LIBeLIS - Java Forum Stuttgart

LIBeLIS
JDO verleiht Flügel
Enterprise Information Access
Scalable, robust JDO solutions
« JDO: Just Do Objects ! »
Alexander v. Zitzewitz
[email protected][email protected]
 www.libelis.com
1
Business Card
Enterprise Information Access Vendor
Founded in 2000
by experts in DBMS and J2EE technologies
Worldwide operations
Direct operations in UK, Germany and France (+Spain)
Distributed in USA & Canada
Investors
First funding round in November 2002
Strong commitment to Java standards
Member of the Java Community Process (JCP)
Active Member of the JDO Expert Group
Charter Member of JDOCentral
Board Member of the ObjectWeb consortium
2
LiDO
LiDO : Flagship product
NAViLIS as a side-product (eBusiness browser)
FileDB a light & fast 100% Java embedded DB
First JDO 1.0 implementation
Enterprise level, scalable, robust JDO implementation
Production ready
Available since July 2001
>25.000 downloads worldwide
Very active community
3
The JDO standard
Java Data Objects
The new specification for Java
object persistence from Sun
4
JDO
The Java standard for data access
Final Approval in March 2002
Main contributors
Sun, IBM, BEA, Oracle, SAP ... & LIBeLIS
+O/R mapping tools vendors + ODBMS vendors
Does not replace but complement JDBC and EJB
Goals
Transparent persistence
Not intrusive in Business Objects
Universal Data Access
RDBMS, ODBMS, binary, XML, text, Legacy, applications, TP monitors,
MOM...
Independant deployment
Java level: embedded J2ME, client-server J2SE, distributed J2EE
Might be embedded within in a future JDK release
Books
5
JDO programming
LiDO from the
developer point of view
Transparent data access with full
Java support
6
LiDO Object model: no restriction
public class Employee extends Person {
public String name;
private java.util.Date birthday;
private int salary;
public Company company;
public IActivity activity;
Vector awards;
public void bonus(int more) {
salary += more;
}
}
No need for any
javax.jdo import
Inheritance
Interfaces
Collections
No need for
accessors /
mutators
This is a persistent class !
If you can write it in Java, any JDO engine should store it.
7
LiDO build cycle
source
code
javac
compiler
byte
code
any JDO
enhancer
java
JVM
Mapping
(xml file)
byte
code
LiDO
Enhancement is fully defined by the JDO
spec.
Enhanced byte-code is portable across JDO
implementations
Enhanced byte-code remains the same for any
data source
Allows debugging and profiling (JPDA)
Allows fully transparent persistence
8
Programming with LiDO
LiDO


Application objects
Some explicit JDO calls:
•Connection(*)
•Query
•Transaction(*)
rdbms
Business objects
Fully Transparent (no JDO calls):
•Navigation between objects
•Data manipulation in Java
•Automatic data mapping
•Client cache, object identity
odbms
Java VM
( *)
: in a J2EE context, connections and transactions
are managed by the application server itself !
9
JDO within an
Enterprise Application
design
LiDO from the
architect point of view
Efficient Enterprise Information
Access
10
How LiDO complements J2EE
JSP
Session
Entity
Java
Entity Bean
LiDO
Does not scale (even CMP 2)
Efficient persistence layer
RDBMS centric
Universal, any data source
Limited object support
Full Java
High deployment costs
Infra-structure independant
High development costs
Straight Java classes
Any
Data
Source
JDO is already
perceived
as one of the
most critical
Java
extensions as it
perfectly
complements
J2EE !
11
LiDO Design
Universal & Scalable
JDO Implementation
« JDO: Just Do Objects ! »
15
LiDO architecture
JDO implementation
Transaction, JDO QL, State Manager
Caches, Dual Query Engine,
JCA Connection pool, ...
Specification implementation
Core technology
Storage Managers
RDBMS
ODBMS
Binary
files
...
Data source dependant layer
The only JDO solution that is more
than just another O/R mapping tool !
16
JDO Demo
LiDO Live !
It doesn’t only run on
PowerPoint 
17
LiDO Delivers
20 to 50% LESS source code to write
Typical number of lines within the business object
models used to store and retrieve these objects: 30% and
this may increase should your model be complex.
Significant Performance Improvements
5 to 40 times faster than EJB CMP
Significant Reduction of complexity
No modelling restrictions
18
Conclusion
Too Good to be True ?
Do not trust us…
just Download and Check it out !
www.libelis.com
EMail: [email protected]
Trust our Customers:
Societe Generale, France Telecom,
Orange, CARDIF, SANOFI, Raytheon,
FGM, Rothschild Bank, CRIL, LineData
Services …
21