Table 1. Cache definitions - IBM System
z9 versus IBM System z10™ test
DynaCache size
Caching mode
Servlet caching
Replication domain
Replication type
20000
Distributed map
Disabled
Entire domain
Not-shared
The test used one LPAR on a 16-way IBM System
z10, equipped with same amount of physical resources as the IBM System z9 LPAR. The guest definitions were
the same, except that in the IBM System z10 environment the
number of virtual CPUs assigned to the Proxy Server guest was increased
from one to two because the single CPU was fully utilized and gating
the throughput. Table 2 described
the test configuration.
Table 2. Guest configuration
for IBM System z10 environment
System/Guest
Memory size
Number of CPUs
z/VM®
1248 MB
8
Caching Proxy Server (CPS)
-
-
WebSphere® Proxy
Server (PS)
750 MB
1/2
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
Figure 1 compares
the throughput for the two systems when the number of workload clients
is scaled from 40 to 70. Figure 1. Throughput
comparison - client scaling, z9® versus z10, WebSphere Proxy Server.
Figure 2 compares
the CPU utilization for the two systems when the number of workload
clients is scaled from 40 to 70. Figure 2. CPU
utilization comparison - client scaling, z9 versus z10, WebSphere Proxy Server.
Observations
Replacing the IBM System
z9 system with an IBM System z10 system resulted
in a throughput increase of 86%. WebSphere cluster
utilization increased by 18%. The WebSphere Proxy
Server utilization increased by 28%, and the DB2 and Firewall utilization showed a slight
increase. Total z/VM utilization
is peaking at 772%.
Conclusions
The IBM System
z10 system obtained a significant throughput advantage over the IBM System z9 system. The Trade workload used
in this test demonstrates the performance advantage of the IBMSystem z10 system.
