Servlet implementation

HPE Service Manager uses a servlet implementation to manage client connections to the server. The servlet implementation uses a pre-configured embedded Java servlet that does not require any additional installation or configuration of Java components. Administrators can manage the servlet implementation using the HPE Service Manager configuration files and management procedures rather than Java interfaces.

A servlet implementation provides the following management features:

• The ability to specify the exact communications ports client connections use
• The ability to specify the maximum number of client connections HPE Service Manager will accept
• Industry-standard resilience and scalability options
• Multithread processing capabilities

In a servlet implementation, administrators specify the communications ports the system uses in advance. The total number of communication ports required for a servlet implementation is based on the number of servlet container processes you want the HPE Service Manager host to support. Each servlet container process manages a set number of client sessions as determined by the number of threads it contains.

Each servlet container process supports a set number of client connections over two communications ports, one for HTTP communications and one for HTTPS communications. Administrators set the number of client connections per process using the threading parameters. Two communications ports is sufficient for up to approximately 150 client connections (as determined by the HPE Service Manager host's memory requirements).

A servlet implementation is ideal for administrators who want to do capacity planning and system resource management. The communications port and threading parameters allow administrators to control the resources that client connections consume. By specifying the communication ports and number of threads available for client connections, administrators can set a limit on the total number of client connections any one hardware resource supports. Setting connection limits prevent client connections from consuming more system resources than desired and allows administrators to set a server load threshold in advance. If a client attempts to connect to a HPE Service Manager system that has reached its connection limit, the server refuses the connection and displays an error message that the server is unavailable.

Certain servlet implementations benefit from the native resilience and scalability features of grouping Java servlets. In a horizontal scaling implementation, administrators can create virtual groups where multiple HPE Service Manager servers act as a single system. Each member in a virtual group communicates its connection availability and system resource usage to the other members of the group. Should a group member fail, the other group members are unaffected by the outage and the system can route any new client connection requests to another member in the virtual group. If the client connection routing process fails, the existing client connections are unaffected and an administrator can restart the client connection routing process.