Oracle Internet Application Server 8i Migrating from Oracle Application Server Release 1.0 A83709-01 |
|
This chapter provides a general discussion of the Oracle Internet Application Server characteristics in comparison to those of Oracle Application Server (OAS). It includes a mapping of OAS components to their equivalent functionality in Oracle Internet Application Server.
Oracle Internet Application Server is a middle-tier application server designed to enable scalability of web and database-centric applications beyond the limits of a single database instance. It offers:
The table below shows OAS components and their corresponding functionality in Oracle Internet Application Server.
Table 1-1 OAS and Oracle Internet Application Server functionality
This section discusses enterprise services, characteristics of a web site of concern to administrators and developers. It describes scalability, availability, fault tolerance, load balancing, and administration in OAS and how they will work after you migrate your site to Oracle Internet Application Server.
OAS consists of the HTTP layer, the server layer, and the application layer. The HTTP listener layer is made up of the HTTP server and the dispatcher. The Server layer provides a common set of components for managing these applications. These components include load balancing, logging, automatic failure recovery, security, directory, and transaction components. The application layer is made up of applications, cartridges, and cartridge servers. When a request arrives, the dispatcher routes the request to the application server layer and if a cartridge instance is available, the request will be serviced by that instance, otherwise a new instance will be created.
Similarly in Oracle Internet Application Server, the Apache Web Server (HTTP server), and mod_jserv run in the same process. Apache JServ is a pure Java servlet engine and runs in a separate process. The Apache Web Server uses mod_jserv to route requests to an Apache Jserv processes, much like the dispatcher in OAS.
OAS can be deployed in single- or multi-node environments. Similarly, the Apache Web Server and Apache JServ can be configured for single or multi-node environments.
In OAS, each listener can accommodate a maximum number of concurrent connections. This number varies based on operating system restrictions. To distribute the request load on a site, you can create multiple listeners, each listening on a different TCP port.
On Unix platforms, Apache Wed Server on start-up, creates a pool of child processes ready to handle incoming client requests. As the requests are processed and the load increases, Apache will make sure that there are a few extra processes running for subsequent requests. The initial and maximum size of the pool, and the min/max number of spare server processes, is configured with the StartServers, MaxClients, MinSpareServers and MaxSpareServers directives respectively.
On Windows NT, Apache runs as a multi-threaded process. The number of simultaneous connections is configured with the ThreadsPerChild directive, which is analogous to both the StartServers and MaxClients directives for UNIX.
Apache Web Server can be configured to run multiple instances on the same host, each of them using a different IP address/TCP port combination, or on different hosts.
In OAS, as the number of requests increases, the system creates new cartridge servers and new instances in them.
In Apache, mod_jserv receives requests from the Apache server and routes them to Apache JServ, the servlet engine.
Apache Jserv runs all servlets within servlet zones. Some of the advantages are: better security, run multiple JVMs and support for multiple virtual hosts.
When a component such as a listener or a cartridge server fails, OAS detects the failure and restarts the failed component, restoring any preserved state information when possible.
In Apache, if there is more than one Apache server host, or more than one JServ host, and one of them stops, the system will still work as long as there is one Apache server and one JServ running. A last known status is maintained for every JServ, and any Apache server can route a request to any Apache JServ.
In Apache, the administrator is responsible for restarting any failed Apache Web Server or Apache Jserv instances.
OAS allocates system resources and prioritizes requests based on two types of load balancing methods: priority-based and min/max.
In priority mode, the system manages and allocates resources automatically, based on the priority level you set for your applications and cartridges. The number of processes, threads, and instances is automatically determined based on the request load and priority level of the application and components.
In min/max mode, you set the number of instances, threads and client parameters for each cartridge at the cartridge level.
In Apache, you define the number of JServ hosts, host weight, and a logical set of these hosts in your configuration file. The system assigns incoming requests to JServs. If a JServ fails, requests are redirected to the other members of the logical set.
OAS provides GUI tools and built-in support for administering and monitoring your site, listeners, and applications.
In Apache, you perform site administration and maintenance by editing the Apache server and Apache JServ configuration files.
In Oracle Internet Application Server, as in OAS, GUI tools are provided for site administration and maintenance. The difference in the number and type of configuration files is significant:
OAS supports a number of different security schemes for both user and host authentication, SSL, and the Oracle Wallet Manager.
In Oracle Internet Application Server, Apache JServ can run behind a firewall (the AJP protocol uses only one TCP port). It uses ACL (allowing AJP requests only from hosts with ACL) and supports SSL.
|
Copyright © 2000 Oracle Corporation. All Rights Reserved. |
|