Abstract: Disclosed are a function of automatically tuning a communication performance between the communications devices connected to a network and a function of periodically confirming and tuning the communication performance. In a plurality of communications devices capable of performing the communications by switching over a plurality of communication modes, the communication performance in each of the communication modes is previously measured under a plurality of communication conditions, there is obtained, per communication condition, a condition-based optimum communication mode as a communication mode in which the communication performance in the specific communication mode exceeds a communication performance in other communication mode, the condition-based optimum communication mode is selected in accordance with the communication condition when in communications, and the communications are thus performed.

Abstract: A storage system includes a cache device that has a cache memory as a high-speed-access storage medium, a disk device that has a disk as a high-capacity storage medium, and a control device that accepts an access request from a client device and makes the cache device and the disk device execute a process corresponding to the access request. The cache device, the disk device, and the control device are decentralized on a network.

Abstract: There is provided a storage system including a storage unit capable of storing therein file data, a plurality of file servers for effecting file processing on the storage unit in response to a file request concerning file data which is received from a client through an external network, a file server administrating node for unitarily administrating transfer processing for transferring the file data to the file server based on the file request and reply processing for sending a reply message concerning the file request to the client, and an internal network for interconnecting the storage unit, the file servers and the file server administrating node so that communication can be effected in the internal network. According to the foregoing arrangement, it is possible for the storage system to have satisfactory scalability capable of coping with the expansion of band of the network at a low cost.

Abstract: A storage system includes a cache device that has a cache memory as a high-speed-access storage medium, a disk device that has a disk as a high-capacity storage medium, and a control device that accepts an access request from a client device and makes the cache device and the disk device execute a process corresponding to the access request. The cache device, the disk device, and the control device are prepared decentrally on a network.

Abstract: There is provided a storage system including a storage unit capable of storing therein file data, a plurality of file servers for effecting file processing on the storage unit in response to a file request concerning file data which is received from a client through an external network, a file server administrating node for unitarily administrating transfer processing for transferring the file data to the file server based on the file request and reply processing for sending a reply message concerning the file request to the client, and an internal network for interconnecting the storage unit, the file servers and the file server administrating node so that communication can be effected in the internal network. According to the foregoing arrangement, it is possible for the storage system to have satisfactory scalability capable of coping with the expansion of band of the network at a low cost.

Abstract: The present invention relates to a multicast transmission method which is capable of realizing a multicast function while maintaining the independence of each of nodes concurrently with minimizing the traffic on a communication network. In the multicast transmission method according to this invention, a data transmitting node, in which a multicast transmission request occurs, adds a multicast address to a data send message and then broadcasts the data send message plus multicast address to all the nodes, while each of the nodes receiving the multicast address refers to its own multicast table to decide whether or not a multicast address coincident with the received multicast address exists in the multicast table.

Abstract: The invention provides an information retrieval system wherein, even when a plurality of data bases are retrieved simultaneously, information retrieval can be performed at a high speed without necessity for special hardware and without being influenced by the magnitudes of the data bases of a retrieval object. The information retrieval system includes a plurality of retrieval servers for performing retrieval processing, and a retrieval management server for managing operation of the plurality of retrieval servers. The retrieval management server is constructed so as to divide a data base of a retrieval object and relating information regarding the data base and allocate resulting divided parts of the data base and the relating information corresponding to the divided parts in sets to some or all of the plurality of retrieval servers.

Abstract: A memory protection setting unit sets a protection area in a multiport memory provided with two ports, that is, port A and port B. The CPU accesses the multiport memory through port A. At this time, an area other than the protection area can be accessed. The CPU transmits data to an external device through a transfer control device. The transfer control device accesses the multiport memory through port B. Only the protection area can be accessed at this time. Upon receipt of a request from the CPU, the transfer control device writes data received from the external device to the protection area and reads data to be sent to the external device from the protection area. Protection is removed by the memory protection setting unit from the protection area where data is read or written by the transfer control device.

Abstract: A disk access apparatus for performing a stride processing of data is used in a parallel computer system for connecting a plurality of computers to each other through a network.

Abstract: In a memory, a program for internal control such as temperature compensation and calibration of a disk device and a program for scheduling the transfer timing and transfer size and effecting the control of data stream transfer between a computer and the disk device according to the scheduling are stored as firmware. A CPU controls a timer interrupting section, disk interface driver and internal interface driver according to the program to attain the internal control of the disk device and data transfer between the computer and the disk device. The timer interrupting section permits data transfer according to the scheduling by previously setting the interrupting timing by the CPU and generating an interruption according to the thus set timing.

Abstract: In a file system for use with a parallel computer, a file access system is provided. When a host computer opens a file, it can see it as one file. When each system opens a file, it can see the file as a local file. A substance is stored as one file in a secondary memory. However, in an open mode, each cell can see a required portion.

Abstract: A disk block controller and file system support large extended files by using special logical blocks which contain a number of sequential physical disk blocks. The disk block controller includes a file creation support unit, an extended management unit, a first and second management units, and an allocation unit. The file creation support unit separates free physical disk blocks into a region with a small number of sequential empty physical disk block spaces and a region with a large number of sequential empty physical disk block spaces, and allocates these regions to different parts of the disk. The extended management unit manages a start and sequential block numbers for sequential empty physical disk blocks having specified entries and is utilized for extended files whose logical block size is an integral multiple of a physical block.

Abstract: An input/output control unit switches modes between a parallel access mode in which a plurality of I/O interfaces, to which a plurality of external storage devices are connected, are collectively controlled as if they were a single interface and an individual access mode in which the plurality of interfaces are individually controlled. Data can be distributed or combined when transmitted between the external storage devices and a network.