When you use multiple threads concurrently, you must generate instances per thread since the class used in the processing of the request message and the response message cannot be shared with multiple threads.

If you generate instances per analysis, the performance degrades, since the instance generation process takes time.

If you use the instance pool function, you can save the instances generated in the instance pool that is created on memory. You can reuse instances by sharing the instances saved in the instance pool across the entire system, thereby improving the processing efficiency.

To save instances generated in the instance pool, perform settings as per the following table:

Table 5-10 Properties used for setting the number of instances cached in an instance pool

Definition file	Property	Contents
HCSC server runtime definition file	xml-pooled-instance-minimum	Minimum number of instances of a class used in the process of request message and response message to be cached
	xml-pooled-instance-maximum	Maximum number of instances of a class used in the process of request message and response message to be cached

When you are reusing already created instances, define the minimum number of caches. If you define the minimum number of caches, the pool is initialized when you start the HCSC server. Generate the instance of the defined value and set it in the pool.

If you set the maximum number of caches, you can save the instance till it reaches the defined value. Even if you reduce the instances by reusing them, you need not perform depletion, since an instance is to be generated and set in the pool.

For details on the xml-pooled-instance-minimum property and the xml-pooled-instance-maximum property, see "HCSC server runtime definition file " in "Service Platform Reference Guide".