摘要:
A data transfer management method and apparatus for reducing collision during data transfer. The invention stores a data length (m) in byte units of a plurality of packets capable of being transferred in a predetermined transmission time interval, stores the predetermined transmission time interval (t), accumulates the quantity of the packets to be transferred on a communication stream in the predetermined transmission time interval, and limits the quantity of the packets to be transferred to a value not greater then the data length (m).
摘要:
A periodic process scheduling method which includes receiving requests for periodical execution of processes each with a period of execution and execution time, securing allocation time for the shortest process to be executed in respective periods of execution, securing another allocation time for part of a process if full execution time of the process cannot be secured within the respective periods of execution, and starting execution of the process in secured allocation time. A data transfer management method for reducing collision in data transfer which includes storing data length of packets being transferred in a predetermined transmission time interval, storing the predetermined transmission time interval, accumulating a quantity of packets to be transferred in the predetermined transmission time interval, and limiting quantity of packets to be transferred to not be greater than data length.
摘要:
A data transfer management method for effectively reducing collision in data transfer includes the steps of storing a data length (M) of a plurality of packets capable of being transferred in a predetermined transmission time interval, storing the predetermined transmission time interval (t), accumulating the quantity of the packets to be transferred on a communication stream in the predetermined transmission time interval, and limiting the quantity of the packets to be transferred to a value not greater than the data length.
摘要:
Periodic process scheduling method of computer system includes steps of receiving requests for periodical execution of processes each with period of execution and execution time, securing allocation time for process having shorter period than the other processes to be executed in each of respective periods of execution for process, securing another allocation time for part of another process within first of respective periods of execution of process other than allocation time in first of the allocation time for process if full execution time of another process cannot be secured within first of respective periods of execution of process, and starting execution of process in secured allocation time and another process in secured another allocation time. A data transfer management method for effectively reducing collision in data transfer includes steps of storing data length (M) of packets capable of being transferred in predetermined transmission time interval, storing predetermined transmission time interval (t), accumulating quantity of packets to be transferred on communication stream in predetermined transmission time interval, and limiting quantity of packets to be transferred to value not greater than data length.
摘要:
A process executing method capable of performing multiprocessing by using a shared resource without impairing periodical drivability of processes designed for executing continuous media processing. When a process requests the use of the shared resource, abortion of that process is first disabled by the process itself by using an abort disable function and then preemption of the same process is disabled by means of a preempt control module, whereupon the process enters a processing executed by using the shared resource. Upon completion of the processing for the shared resource, the process is immediately set to a preempt-enabled state by means of a preempt control module. After completion of all the processings, the abort-disabled state is finally cleared by using a disabled-abort clear function. Upon occurrence of forcive termination of a process in the abort-disabled state thereof, execution of this process is continued until the abort-disabled state is cleared, and the process is terminated forcibly after the abort-disabled state has been cleared.
摘要:
A process executing method capable of performing multiprocessing by using a shared resource without impairing periodical drivability of processes designed for executing continuous media processing. When a process requests the use of the shared resource, abortion of that process is first disabled by the process itself by using an abort disable function and then preemption of the same process is disabled by means of a preempt control module, whereupon the process enters a processing executed by using the shared resource. Upon completion of the processing for the shared resource, the process is immediately set to a preempt-enabled state by means of a preempt control module. After completion of all the processings, the abort-disabled state is finally cleared by using a disabled-abort clear function. Upon occurrence of forcive termination of a process in the abort-disabled state thereof, execution of this process is continued until the abort-disabled state is cleared, and the process is terminated forcibly after the abort-disabled state has been cleared.
摘要:
A scheduling method of a periodic process of a computer system for keeping the execution interval of each process group constant as far as possible in the case where a plurality of process groups including periodically executed processes are executed in parallel. If a group master process of each process group specifies a wakeup interval period and a required CPU time per period and requests allocation of a CPU time, then a CPU allocation time of a specified process group is secured so as not to collide with a CPU allocation time of another process group, and a scheduling table is created so as to maintain the specified wakeup interval period. In response to arrival of time when one of the process groups should be waked up, a kernel process or a scheduler conducts wakeup by changing execution priority of a process belonging to this process group to highest priority ("raised") in the system, and thereafter maintains the highest priority for CPU allocation time consecutively allocated.
摘要:
Before starting data transfer, the sender node sends the CONNECT message requesting security of hardware resources necessary for data transfer to the receiver node. Each node relaying the CONNECT message secures the CPU time and bandwidth necessary for data transfer and sends the CONNECT message to the receiver node. Upon receipt of the CONNECT message, the receiver node reserves the CPU time necessary for data reception and then sends the ACCEPT command to the sender node. The sender node receives the ACCEPT command and then sends a data packet.
摘要:
A storage media and a storage area of a storage apparatus are associated, and the storage media is used to manage the contents of the storage apparatus. A storage media for storing files is detachably mounted on an information processing apparatus, the information processing apparatus is connected to a storage apparatus via the Internet, the storage apparatus stores files corresponding to the files stored in the storage media, the storage media and the storage apparatus authenticate each other via the Internet, and the information processing apparatus reads files from or writes files into the storage media or the storage apparatus on the condition that the foregoing authentication is successful.
摘要:
A new storage tier is added to a plurality of storage tiers configured in storage areas of a plurality of types of storage devices, whose processing performance is different from each other, without changing its operation form.If a new Tier3 constituting a file system FS4 is added between Tier2 and Tier3, a file server 16 adds a registered generation 1 to a tier number change management table 78, sets Tier3 as the added file system FS4 in the registered generation 1, manages Tier4 as a file system FS5, which has been changed from the former Tier3, adds a policy corresponding to a policy number 1005 to a migration policy management table76, updates only the registered generation from 0 to 1 for policy numbers 1002, 1005, and manages the migration policy management table.