WebSphere Application Server V8.5 Administration and Configuration Guide for the Full Profile

538 WebSphere Application Server V8.5 Administration and Configuration Guide for the Full Profile

15.2.3 Workload management benefits

Workload management provides the following benefits to WebSphere applications:

 It balances client processing requests, allowing incoming work requests to be distributed

according to a configured WLM selection policy.

 It provides failover capability by redirecting client requests to a running server when one or

more servers are unavailable. This redirection improves the availability of applications and

administrative services.

 It enables systems to be scaled up to serve a higher client load than provided by the basic

configuration. With clusters and cluster members, additional instances of servers can be

added easily to the configuration.

 It enables servers to be maintained and upgraded transparently while applications remain

available for users.

 It centralizes administration of application servers and other objects.

15.3 High availability and failover

High availability is also known as resiliency. High availability is the description of the system’s

ability to tolerate a certain amount of failures and to remain operational. This section provides

information about WebSphere Application Server high availability concepts and features.

15.3.1 Overview

High availability (HA) means that your infrastructure continues to respond to client requests

no matter what the circumstances are. Depending on the errors or failures, the infrastructure

can run in a degraded mode. HA is achieved by adding redundancy in your infrastructure to

support the system when failures occur. Availability impacts both performance and scalability.

Depending on your needs, you have to define the level of HA for your infrastructure.

The most common method of describing availability is by the “nines” or the percentage

availability for the system. For example, 99.9% of system availability represents 8.76 hours of

outage in a single year. Table 15-2 shows the level of availability and the calculated downtime

per year.

Table 15-2 Availability matrix

You can calculate availability using the following formula, where MTBF is the mean time

between failure and MTTR is the maximum time to recovery:

Availability = (MTBF/(MTBF + MTTR)) X 100

Keep in mind that the overall infrastructure is available only if all the components are

available. For a WebSphere infrastructure that is composed of several components, such as

Availability % Downtime per year

99% (two 9s) 87.6 hours

99.9% (three 9s) 8.76 hours

99.99% (four 9s) 56.56 minutes

99.999% (five 9s) 315.36 seconds

Chapter 15. Clustering, workload management, and high availability 539

load balancers, HTTP servers, application servers, database servers, and so on, availability is

determined by the weakest component.

For most of the environment’s components, several degrees of HA implementation exist. The

cost of the infrastructure is directly linked to the level of availability. Evaluate the business loss

of the infrastructure downtime, and ensure that the business case justifies the costs. Moving a

system availability from 99.9% to 99.99% can be expensive. It can also be true that the

system is used only during regular business hours on regular working days. This use implies

that an availability of 99.9% is more than adequate to meet the operational window.

For additional information about this topic, go to the following website:

http://www.ibm.com/developerworks/websphere/techjournal/0312_polozoff/polozoff.htm

l#sec1

Because it is likely that the complete environment is made up of multiple systems, the goal is

to make the whole system as available as possible by minimizing the number of single points

of failure (SPOFs) throughout the system by adding redundancy. Redundancy can be added

at different layer, such as hardware, process, and data.

15.3.2 WebSphere Application Server high availability and failover

This section provides information about the WebSphere Application Server HA features. It

can help you to understand how the HA features work and can assist you in planning and

configuring for HA.

Figure 15-15 represents a typical WebSphere Application Server topology, where a request

travels through the load balancer to the HTTP server to the web container and finally to the

EJB container.

Figure 15-15 WebSphere topology and incoming request

The sections that follow provide information about HA and failover options for components

and services that are involved in the processing of the request.

Load balancing

You can configure the Edge Components Load Balancer to run as an active-passive pair. The

active instance is the single point of distribution for HTTP requests. If it fails, the passive

instance routes the requests.

Java

client

Name

service

Plug-in

Load balancer

Plug-in

configuration

HTTP server

Routing

table

WebSphere

Application Server

Web

container

WebSphere

Application Server

EJB

container

Routing

table

IIOPHTTP(S)HTTP(S)