Unless otherwise noted, this is the standard test environment
for the WebSphere® Application
Server tests.
Figure 1 illustrates
the test environment. See Test environment for more details.Figure 1. WebSphere Application
Server test environment.
There were a total of four WebSphere Application
Servers in the cluster, one workload simulator machine, one Apache
Web Server and one WebSphere Edge
Component Proxy Server or WebSphere Application
Server Proxy Server node.
Table 1 shows the setup
of the standard configuration.
Table 1. Setup
of standard configuration for the WebSphere Application
Server tests.
System/Guest
Memory size
Number of CPUs
z/VM®
12288 MB
8
Caching Proxy Server
512 MB
1
Apache HTTP Server
512 MB
1
WebSphere Application
Server 1
2 GB
2
WebSphere Application
Server 2
2 GB
2
WebSphere Application
Server 3
2 GB
2
WebSphere Application
Server 4
2 GB
2
DB2® Server
4 GB
1
Firewall 1
512 MB
1
Firewall 2
512 MB
1
The Trade application was configured for 1000 stock quotes and
2000 users.
Test case scenarios
These are the test case
scenarios that were run:
Baseline definition
This test case uses the four node cluster with Distributed Map
mode. The number of workload generating clients were scaled from 25
to 80 to define the base line. See Baseline definition parameters and results.
Vary caching modes
This test case uses the three Trade caching options: No Caching,
Command Caching, and Distributed Map caching. For Command Caching,
an additional scenario was done with Servlet caching enabled. The
objective was to determine which caching mode produced the highest
throughput. See Vary caching mode.
Vary garbage collection policy
This test case uses measured the impact of the four possible JVM
garbage collection policies that can be set for the JVM. See Vary garbage collection policy.
Vary caching policy
This series of test cases compare the various cache replication
sharing policies. These were done for the four sharing types: not-shared,
shared-push, shared-pull, shared-push-pull. They were run with the
Trade caching option set to Distributed Map. See Vary caching policy.
Vary proxy system in the DMZ
This test case compares the throughput obtained with the WebSphere Edge Services Caching
Proxy Server, the WebSphere Proxy
Server, and no proxy server.
With no proxy server, the Apache Web
server was moved into the DMZ and the Trade workload simulator was
configured to direct all request to the Apache Web server. See Vary proxy system in the DMZ.
Scaling the size of the DynaCache
This series of test cases compare the various dynamic cache sizes.
They were run with the Trade caching option set to Distributed Map.
The dynamic cache size was varied from 2000 to 20000. See Scaling the size of the DynaCache.
Scaling the update part of the Trade workload
Updates from data are the critical case for caching, because the
caching mechanism has to make sure that data is consistent system-wide
in all caches and variables. In this test case, the update part of
the Trade workload was scaled. Normally 14% of the transactions in
the Trade workload require a data base update. In this test case,
the data base update activity was scaled from 0% to 14% in increments
of 2%. See Scaling the update part of the Trade workload.
