Vary garbage collection policy

Edit online

The impact of the possible JVM garbage collection policies were tested in this scenario.

Table 1 describes the four possible JVM garbage collection policies that can be set for the JVM.

Table 1. Garbage collection policies for JVM
Policy Option Description
Optimize for throughput Xgcpolicy:optthruput (optional) This is the default policy. It is typically used for applications where overall throughput is more important than short garbage collection pauses. The application is stopped each time that garbage is collected.
Optimize for pause time Xgcpolicy:optavgpause Trades off higher throughput for shorter garbage collection pauses by performing some of the garbage collection concurrently. The application is paused for shorter periods.
Generational concurrent Xgcpolicy:gencon Handles short-lived objects differently than objects that are long-lived. Applications that have many short-lived objects can see shorter pause times with this policy, while still producing good throughput.
Subpooling Xgcpolicy:gencon Uses an algorithm similar to the default policy's, but uses an allocation strategy that is more suitable for multiprocessor machines. This policy is of benefit to SMP machines with 16 or more processors. This policy is available only on IBM®System p® and IBM System z® platforms. Applications that need to scale on large machines can benefit from this policy.

This test case was run to see if changing the JVM's garbage collection policy would have any effect on the throughput. The workload driver was set to 30 clients for all runs and Trade was set to Distributed Map caching. The standard configuration with WebSphere® Edge Services Caching Proxy Server was used.

All runs were done with the cache configuration described in Table 2.
Table 2. Cache definitions - Vary garbage collection policy
DynaCache size Caching mode Servlet caching Replication domain Replication type
20000 Distributed Map Disabled Entire domain Not shared
Figure 1 illustrates the throughput using the various garbage collection policies.
Figure 1. Garbage collection policy comparison
This figure shows the throughput results for the garbage collection policy comparison.

Observations

The default garbage collection policy of optimize for throughput resulted in the highest throughput. The other garbage collection policies resulted in lower throughput with optimize for time having the lowest throughput.

Conclusion

For the Trade workload, the default garbage collection policy produced the best throughput.