Application Server 10g Load balancing.
Documents Created by : Bharat Gali
Created date : 01-Feb-2006
Application Server Version :10.1.2.0.2
Last Revised :30-Apr-2006
*********************************************************************************
This document focused on Load balancing concepts with some examples that I implemented with some of my clients and did n't cover the basic concepts of Application server. For more information Application server and its concepts please refer Oracle site.
1) Process death detection and automatic restart.
OPMN tracks any application server components (Forms,Report,Portal etc etc) if it is register with it. OPMN continuous pings its registered components and writes to logs about the state and status.
OPMN detects Oracle Application Server processes that are down or unreachable, it automatically restarts them.
Example OPMN command
$ORACLE_HOME/opmn/bin opmnctl status
2)Active-Active Load balancing.
Two or more Oracle Application Server instances are configured to serve the same workload. These instances can run on the same machine or on different machines. The instances are front-ended by an external load balancer, which directs requests to any of the active instances. Instead of an external load balancer, you can also run a software load balancer to distribute the requests. In production environment, however, a hardware load balancer is recommended.
Advantages of an active-active topology:
Run OC4J instances on multiple nodes.
OC4J clustering is very important in application server load balancing to achieve maximum service availability. Below link provides implementation techniques of OC4J clustering
Click here for implementation techniques of OC4J clustering with Examples
e
High Availability Model1 Via Web cache As Shown in the above figure.
Hardware load balancer is connected to 2 web cache servers
1) Web Cache1
2) Web Cache2.
Based on Availability and load web cache servers direct the request to either M1 OR M2 SERVER.
Middle Tier servers M1 and M2 are connected to two infra structure services which share common OID repository.
Disadvantages of this model
There is no redirection of the session for OCJ4 with this model. For examples Some one already connected to OCJ4 application (Forms, reports and ETC) and there is a failure in the service. With this model there is not re direction mechanism to the second OC4J. This can be achieved thought OC4J clustering. Which is shown in the next model
High Availability Model2
The above model is same as model 1 except it is having OC4J clustering. OC4J clustering can handing the re-directions which are failure in the middle
Example
If the application is connect via OC4J 1 and failed. With the OC4J clustering it is still possible to transfer to second OC4J.
Below is the link for OC4J clustering implementation.
OCJ4 clustering.
Above model there are no Web cache servers. Load balancing routers is directly connected to OC4J Servers. And only one infrastructure instance is running. Normally the failure in infrastructure are very less compare to the failures of middle tier. This is the typical model which is implemented lot of place.