Process management apparatus, computer systems, distributed processing method, and computer program for avoiding overhead in a process management device

A computer system comprising:

• a plurality of processing devices each of which is configured to execute a requested task and to transfer the executed result to a destination designated by the requester of the task;
• and
• a process management apparatus connected to the processing device via an internal bus, wherein the processing device and the process management apparatus are connected to a network via the internal bus, respectively, the process management apparatus comprising:
• first management means for allowing access to a predetermined memory in which a cluster list and communication setting information for enabling network communication with each of the clusters are listed, wherein the processing devices are sorted into a plurality of clusters in the cluster list, and the number of sorted processing devices corresponds to the size or kind of a previously estimated task, wherein, in the cluster list, each cluster comprising at least one processing device which is/are selected based on a combination of the kind of the task that can be performed in each processing device and a processing capability of the each processing device, and the selection is performed per each kind or size of the task which is estimated to be performed, and wherein, in the cluster list, identification information for each cluster representing whether the each cluster can be assigned or not, is recorded;
• assigning means for assigning, upon receiving a task demand from any of the processing device through network communication, the task to a specified cluster which is specified by referring the cluster list for a suitable cluster for the size or the kind of the task;
• and
• second management means for transferring, to the cluster to which the task is assigned by the assigning means or an original processing device that requested the task demand, the communication setting information for the other processing device or other cluster, thereby allowing direct transfer of the executed result between the cluster and the original processing device through network communication.

A device for distributed processing management adapted to be installed in a computer system adapted to be connected to a network to which a plurality of processing devices are connected, each of the processing devices configured to execute a requested task and to transfer the executed result to a destination designated by the requester of the task, wherein the device for distributed processing management executes a predetermined computer program to cause the computer system to operate as:

• first management means for allowing access to a predetermined memory in which a cluster list and communication setting information for enabling network communication with each of the clusters are listed, wherein the processing devices are sorted into a plurality of clusters in the cluster list, and the number of sorted processing devices corresponds to the size or kind of a previously estimated task, wherein, in the cluster list, each cluster comprising at least one processing device which is/are selected based on a combination of the kind of the task that can be performed in each processing device and a processing capability of the each processing device and the selection is performed per each kind or size of the task which is estimated to be performed, and wherein, in the cluster list, identification information for each cluster representing whether the each cluster can be assigned or not, is recorded;
• assigning means for assigning, upon receiving a task demand from any of the processing device through network communication, the task to a specified cluster that is specified by referring the cluster list for a suitable cluster for the size or the kind of the task;
• and
• second management means for transferring, to the cluster to which the task is assigned by the assigning means or an original processing device that requested the task demand, the communication setting information for the other processing device or other cluster, thereby allowing direct transfer of the executed result between the cluster and the original processing devices through network communication.

A method for performing distributed processing in cooperation between a plurality of processing devices and a process management apparatus, each of the processing devices has a function for executing a requested task and transferring the executed result to a destination designated by the requester of the task, and the process management apparatus communicates with each of the processing device via a network, comprising the steps of:

• previously sorting, among a plurality of processing units that are available for communication, the processing devices, the number of which corresponds to the size or kind of the tasks expected to be requested, into a plurality of clusters, and storing the cluster thereof in a cluster list of a predetermined memory area, wherein, in the cluster list, each cluster comprising at least one processing device which is/are selected based on a combination of the kind of the task that can be performed in each processing device and a processing capability of the each processing device, and the selection is performed per each kind or size of the task which is estimated to be performed, and wherein, in the cluster list, identification information for each cluster representing whether the each cluster can be assigned or not is recorded;
• assigning, upon receiving a task demand from any of the processing device through network communication, the task to a specified cluster that is specified from the cluster list and is suitable for the size or the kind of the task;
• and
• transferring, to the cluster to which the task is assigned by the assigning means or an original processing device that requested the task demand, the communication setting information for the other processing device or other cluster, thereby allowing direct transfer of the executed result between the cluster and the original processing device through network communication;
• thereby allowing direct transfer of the executed result of the task without intervening the process management apparatus.

A non-transitory computer-readable medium storing a computer program for causing a computer to operate as a process management apparatus adapted to be connected to a network to which a plurality of processing devices are connected, each of the processing devices configured to execute a requested task and to transfer the executed result to a destination designated by the requester of the task, wherein the computer program causes the computer to function as:

• first management means for allowing access to a predetermined memory in which a cluster list and communication setting information for enabling network communication with each of the clusters are listed, wherein the processing devices are sorted into a plurality of clusters in the cluster list, and the number of sorted processing devices corresponds to the size or kind of a previously estimated task, wherein, in the cluster list, each cluster comprising at least one processing device which is/are selected based on a combination of the kind of the task that can be performed in each processing device and a processing capability of the each processing device, and the selection is performed per each kind or size of the task which is estimated to be performed, and wherein, in the cluster list, identification information for each cluster representing whether the each cluster can be assigned or not, is recorded;
• and
• assigning means for assigning, upon receiving a task demand from any of the processing device through network communication, the task to a specified cluster that is specified by referring the cluster list for a suitable cluster for the size or the kind of the task;
• and
• second management means for transferring, to the cluster to which the task is assigned by the assigning means or an original processing device that requested the task demand, the communication setting information for the other processing device or other cluster, thereby allowing direct transfer of the executed result between the cluster and the original processing devices through network communication.