![[Contents]](FIGURE/CONTENT.GIF)
uCosminexus Service Platform, Overview
![[Glossary]](FIGURE/GLOSS.GIF)
![[Index]](FIGURE/INDEX.GIF)
![[Back]](FIGURE/FRONT.GIF)
![[Next]](FIGURE/AFTER.GIF)
If the multiplicity is incorrectly set, a performance bottleneck might occur.
The following figure shows an example of a bottleneck caused by the multiplicity.
Figure 5-1 Example of a bottleneck caused by the multiplicity
![[Figure]](FIGURE/ZC060100.GIF)
As shown in Figure 5-1 Example of a bottleneck caused by the multiplicity, when a small value is set for the multiplicity for request reception (standard reception or user-defined reception), even if a large number of requests are received, they cannot be processed at the same time. For this reason, the number of processes that can be executed at the same time decreases. Furthermore, even when a higher value is set for the multiplicity for request reception, if the multiplicity defined by the service adapter is small, the service adapter will not be able to execute many processes at the same time. As a result, the service adapter causes a bottleneck and optimum performance will not be achieved.
For synchronous (Web Services) standard reception, multiplicity is set in terms of the number of threads and the maximum number of simultaneous executions. For synchronous (Session Bean) standard reception, multiplicity is set in terms of the number of instances. For details about the multiplicity settings of the HCSC server, see Table 5-1 Properties for setting the multiplicity of standard reception, Table 5-2 Property for setting the multiplicity of user-defined reception, and Table 5-3 Item for setting the multiplicity of a service adapter.
Table 5-1 Properties for setting the multiplicity of standard reception
| Definition file | Property | Description |
|---|---|---|
| HCSC server runtime definition file | request-ejb.instance.minimum | Minimum number of instances for synchronous (Session Bean) standard reception |
| request-ejb.instance.maximum | Maximum number of instances for synchronous (Session Bean) standard reception | |
| request-ejb.parallel.count | Number of threads to be provided for CTM to call applications | |
| request-soap.instance.minimum | Minimum number of simultaneous executions for synchronous (Web Services or SOAP1.1) standard reception | |
| request-soap.instance.maximum | Maximum number of simultaneous executions for synchronous (Web Services or SOAP1.1) standard reception | |
| request-soap1_2.instance.minimum | Minimum number of simultaneous executions for synchronous (Web Services or SOAP1.2) standard reception | |
| request-soap1_2.instance.maximum | Maximum number of simultaneous executions for synchronous (Web Services or SOAP1.2) standard reception | |
| request-jms.instance.maximum | Maximum number of instances for asynchronous (MDB (WS-R)) standard reception |
Table 5-2 Property for setting the multiplicity of user-defined reception
| Definition file | Property | Description |
|---|---|---|
| User-defined reception runtime definition file | user-defined-reception-soap.threads.maximum | Maximum number of simultaneous executions |
Table 5-3 Item for setting the multiplicity of a service adapter
| Window | Description |
|---|---|
| Service Adapter Settings window in the development environment | Maximum number of instances |
For Web Services (SOAP communication), in addition to the number of simultaneous executions, you can set a value for the maximum number of requests in a queue at the Web container level. If you specify the maximum number of simultaneous executions and the maximum number of requests in a queue, you can control the flow when receiving requests from the service requester. This is because you can make a process wait without executing the actual process. The following figure shows the relationship between the requests from a service requester and the maximum number of requests in a queue.
Figure 5-2 Relationship between the requests from a service requester and the maximum number of requests in a queue
![[Figure]](FIGURE/ZC060200.GIF)
For example, Figure 5-2 Relationship between the requests from a service requester and the maximum number of requests in a queue indicates that the maximum number of simultaneous executions is set to 3 and the maximum number of requests in a queue is set to 10. In this case, of the requests that are received simultaneously, no more than three service component invocation requests can be processed simultaneously. If service component invocation requests exceeding the maximum number of simultaneous executions (four or more requests) are received, a number of requests not exceeding the maximum number of requests that can be held in the queue (10 requests) are placed in the pending state. Then, processes will be executed sequentially as soon as the previous process being executed ends.
When the number of requests in the queue reaches the upper limit (10 requests), new requests from the service requester cannot be accepted. Therefore, an error response will be sent to the service requester.
The maximum number of requests that can be held in the queue of the HCSC server is configured by using the properties of the definition files described in Table 5-4 Properties for configuring the maximum number of requests in the queue for standard reception and Table 5-5 Properties for configuring the maximum number of requests in the queue for user-defined reception.
Table 5-4 Properties for configuring the maximum number of requests in a queue for standard reception
| Definition file | Property | Description |
|---|---|---|
| HCSC server runtime definition file | request-soap.queue-size | Size of the pending queue for synchronous (Web Services or SOAP1.1) standard reception |
| request-soap1_2.queue-size | Size of the pending queue for synchronous (Web Services or SOAP1.2) standard reception |
Table 5-5 Properties for configuring the maximum number of requests in a queue for user-defined reception
| Definition file | Property | Description |
|---|---|---|
| User-defined reception runtime definition file | user-defined-reception-soap.queue-size | Size of the pending queue |
All Rights Reserved. Copyright (C) 2015, Hitachi, Ltd.