NonStop Systems Introduction

The Application Server Environment
NonStop Systems Introduction— 527825-001
3-16
Server Handling
The application server components manage the communications links between the
clients and servers and between the servers on different nodes.
Clustering for Fault Tolerance and Scalability
Some application servers, such as WebLogic Server, enable applications to execute as
clusters. A cluster is a group of servers acting as a single server. The cluster appears
to clients as a single server but, in fact, provides the power of several servers working
together. Clustering enhances both the scalability and availability of the application.
Clustering provides scalability because the server’s workload can be dynamically
spread among the other servers in the cluster in response to increasing demand. This
load balancing occurs automatically and is completely transparent to end users. For
even greater performance, the system administrator can add new server instances to
the cluster as business needs dictate.
A cluster uses the redundancy of multiple servers to protect clients from system
failures. The same service can be provided on multiple servers in the cluster, so that if
one server fails, another server can take over the failed server’s workload. As with load
balancing, this failover is completely transparent to the end user. Figure 3-12 on
page 3-17 illustrates the clustering concept. The figure shows how clustered servers
can be spread across multiple processors in a single node and across multiple nodes
in a network.
Figure 3-11. Distributed Processing
Processor 1
Database
C
Database
B
Server C
Server B
Server A
Database
A
VST026.vsd
Process and
link
management
services
Network node
Client
apps
Client
apps
Process and
link
management
services
Network node
Processor 2
Processor 3