Application Servers

1
Application Servers

• Espen Kjölberg
• espen.kjolberg_at_iconmedialab.se

2
Application Servers ???
Personalization
Relibility
J2EE
Components
n-tier
Wap gateway
Business Logic
Web server
CORBA
COM/DCOM
Container
Enterprise Java Beans
Middleware
MTS
Scalable
3
History
70s - 80s Mainframe
80s PC
Late 80s Client Server
90s n-tier (3-tier)
90s n-tier (3-tier)
Mid 90s WWW - 1st gen
Mid 90s WWW - 1st gen
Later 90s WWW - 2nd gen
4
WWW - 1st Generation
DB
5
Problems with 1st gen WWW

• Scalability
• Reusability
• Reliability
• Interacting with other platforms
• Performance
• Security

6
Origin
Application Servers emerged from the 3-tier
solution powered by the evolution of 1st to 2nd
generation Internet.
7
3-tier
Client
Business Components
Business Components
Data
8
Application Servers

• Application Servers in general provide the
  following
• Client connectivity
• Business component connectivity
• Data Source Connectivity
• Transaction
• Security
• Management
• In addition usually support for fail-over and
  redundancy.
• Different application servers usually focus on
  some of the above in depth.
• In Java support for all above is in J2EE.

9
Client Connectivity

• Main issue Increasing number of clients
• Usually solved by the following
• Load balancing of incoming requests
• Dispatch -Using one or several pools of threads
  where a thread corresponds to a request.
• Server clustering - Increase the number of
  servers by creating a cluster, that is the
  clients sees only one virtual server.
• In Java
• JSP
• Servlets

10
Business Component Connectivity

• Possible to interact with other applications and
  platforms, but even more important Used for
  layer separation within an application.
• Examples of Business Logic components
• EJB - Enterprise Java Beans
• CORBA Containers
• COM/DCOM/COM
• (Naming service important part of the components
  which gives the opportunity to run on multiple
  servers.)
• In Java
• RMI Remote Method Invocation
• EJB Enterprise Java Beans
• IDL - ???

11
Data Source Connectivity

• Support any data source connection
• Performance issues handled by
• Database pooling
• Caching
• Request multiplexing
• In Java
• JDBC Java Database Base Connectivity
• JMS Java Messaging Sevice
• JNDI Java Nameing and Directory Interface

12
Transaction

• Standards
• OTS - Corba, JTS - Java interface to OTS
• Two-phase commit
• maintain data in several databases
  simultaneously.
• Transaction Resource
• support participation in two-phase commit trans.
• Transaction Coordinator
• Coordinate two-phase commit trans.

13
Security

• Authentication
• Verify users who wants a connection through a
  client.
• Encryption
• Encrypt sensitive data between clients and server
  or between several servers.
• Access control
• May use the application severs ACL to find access
  rights.

14
Management

• Deployment configuration
• Objects may be used by several appl servers
• Runtime monitoring
• IO, CPU, Memory
• Important because only the jvm is visible from
  the OS, and usually the jvm runs within the
  application server.
• Runtime management
• Specifying Cache, pool sizes and other properties
• Specifying logs
• Defining Servlets
• Defining Enterprise Java Beans

15
WWW - 2nd Generation
Business
Client
Data
16
WWW - 2nd Generation - part II
HTML
WML
17
Another view
18
Another view II
19
Some products

• Apple WebObjects
• ATG - Dynamo
• BEA - WebLogic
• Bluestone - Sapphire/Web
• GemStone
• IBM - WebSphere
• Iona Technologies - iPortal Application Server
• Lotus - Lotus Domino Application Server
• Microsoft - MTS
• Netscape - iPlanet Application Server
• Oracle - Internet Application Server

• ProSyst -Enterprise Java Beans Server Champion
• Silverstream Software - Silverstream
• SoftQuad - HoTMetal Application Server
• SUN - NetDynamics
• Sybase - Enterprise Application Server
• Symantec - Visual Café for Java Enterprise Suite
• Unify Corporation - Unify VISION
• Wall Data - Cyberprise Server

20
IBM WebSphere

• Editions
• Standard
• Java Virtual Machine with support to Java
  Servlets, JavaBeans and JavaServer Pages
• Database Pooling
• Analysis and Monitoring tools
• IBM HTTP Server and support to other web servers
• Advanced
• Multiple machine environment
• EJB and EJB replication support
• Enterprise
• Component Broker
• TXSeries (CICS, Encina)
• Usually used in site with only IBM products, well
  supported in Visual Age, IBMs IDE. One of the
  first on the market.
• Used by Föreningssparbanken, Eurofootball.com

21
BEA WebLogic

• Product Components
• WebLogic Server
• Weblogic clustor for redundancy, fail-over and
  load balancing
• Enterprise JavaBeans Support and MS COM
• Scalable RMI and CORBA Support
• XML Support
• WML and WAP support
• Distributed Transaction support through JTA
• WebLogic Enterprise
• Comprehensive transaction management system
• Mainframe connectivity Support
• BEA Tuxedo application hosting (ATMI)
• Weblogic Commerce Server
• Templates
• Content management integration with Interwoven
  and Documentum
• Webflow
• Usually used in transaction intensive
  applications.
• Well specified integrations with Tuxedo

22
ATG Dynamo

• Product components
• Application Server
• Commerce Server
• Personalization Server
• Scenario Server
• Dynamo Control Center
• Very often used in e-commerce sites
• Full J2EE support, but has several proprietary
  functions.
• Fail-over and redundancy support
• Lot of out of the box functionality
• Used by LetsBuyIt, MsFreckles

23
Oracle Internet Application Server (Oracle 9i)

• Product components
• Java
• Caching
• Portal
• Wireless
• Often used in large Oracle environments
• Full J2EE support, but has several proprietary
  functions.
• Relies on their well-known database
• Used by WideEyes

24
Lessons learned

• In general gives a lot of functionality, which
  you would need to implement yourself.
• Development with application servers implies a
  need of extensive knowledge of the actual
  application server
• Even though Java is used and it should be easy
  to change to another application server, the
  application servers have several proprietary
  functions.
• Observe functions as load balancing, fail-over
  which is hard to implement