Test environment
The test environment used for the WebSphere® Application Server JVM stacking tests consisted of IBM® zEnterprise™ 196 hardware and various software.
Hardware
The tests were done on an IBM zEnterprise 196 (z196) processor. Table 1 lists
the hardware used.
| System | Operating System | Number of CPUs | Memory size | Network | Architecture |
|---|---|---|---|---|---|
| One LPAR | z/VM® version 6.1 | 24 | 200 GB |
|
IBM z196 |
| 1 - 200 Linux™ guests | SUSE Linux SLES 11 SP1 | 24 - 1 Virtual CPUs | 200 - 1 GB | 2 VSWITCH | z/VM guest |
| - | 2 GB | DCSS | |||
| Total | |||||
| IBM zEnterprise 196 | physical: | 24 | 200 GB | Two 1 GB Ethernet | IBM System z® LPAR |
| 4 workload generators | Red Hat Enterprise Linux AS release 4 U3 | 4 | 4 GB | One 1 GB Ethernet | IBM System x® |
The test environment was one LPAR on an IBM zEnterprise 196
with:
- 24 CPUs
- 200 GB memory
- 8 FICON® channels
- Two 1 Gb OSA Express® 3 connections
Network setup
All guests and clients are connected to two VSWITCHs. Each VSWITCH is assigned one OSA port.
The workload drivers were evenly divided between VSWITCHs. Half the drivers were attached to half the WebSphere Servers through VSWITCH1, and the other half through VSWITCH2.
Storage server
The z/VM and Linux guest systems used in this study were allocated in an IBM System Storage® DS8000®, which used the IBM ECKD™ format and FICON channels.
Software
Table 2 lists
the software used to test WebSphere Application
Server horizontal versus vertical stacking.
| Product | Version and release | Comments |
|---|---|---|
| SOA-based workload generator | IBM internal | |
| WebSphere Application Server | 7.1 | Latest patch level |
| SUSE Linux | SLES 11 SP1 | Latest patch level, (2.6.32.27-0.2.2.s390x.rpm) |
| z/VM | 5.4 RSU 0902 | Latest RSU |