JP5686259B2 - Reliable device high reliability apparatus, high reliability method, and high reliability program - Google Patents

Description

  The present invention relates to an apparatus including a plurality of reconfigurable devices such as an FPGA (Field Programmable Gate Array), and more particularly, to an apparatus for achieving high reliability, a high reliability method, and a high reliability program.

  It is very difficult to detect a failure that occurs in a reconfigurable device in an apparatus including a plurality of reconfigurable devices whose hardware configuration can be freely changed according to given configuration information represented by FPGA. The main reason for this is that the area efficiency of a single reconfigurable device is poor, and therefore, when the conventional high-reliability technology is applied, the cost is greatly increased.

  FIG. 16 is a diagram showing a configuration of a conventional high reliability technology (see Patent Document 1). In the example illustrated in FIG. 16, two processors 5100 and 5200 are connected to the comparison unit 5300. That is, the same process is executed in parallel on two processors, and a failure is detected by comparing the execution results for each cycle.

  There is also a technique for testing a configurable integrated circuit using test data as described in Patent Document 2. In this technology, the test data confirms that the desired power consumption and operation speed are obtained when the circuit settings are applied to the reconfigurable integrated circuit and operated. Consists of vectors (consisting of input and output vectors). According to this, whether the circuit setting satisfies the desired operating speed can be determined by the validity of the output vector when the reconfigurable integrated circuit is operated at the desired clock frequency and the test vector is input. It becomes possible.

JP 2006-302289 A JP 2008-004788 A

  However, the method described in Patent Document 1 shown in FIG. 16 has the following problems.

  If this technology is applied to an apparatus having a plurality of reconfigurable devices, all reconfigurable devices having a large area will be duplicated even if a single device is used, resulting in a significant increase in cost. .

  Similarly, the technique described in Patent Document 2 is also useful as a test method for a single reconfigurable device, but is a technique that takes into account the benefits for improving the reliability of a plurality of reconfigurable devices. is not.

  As described above, in an apparatus including a plurality of reconfigurable devices, it is very difficult to detect a failure of the reconfigurable device at low cost.

  Therefore, the present invention provides a highly reliable device, a highly reliable method, and a highly reliable program for a reconfigurable device capable of detecting a failure of the reconfigurable device at a low cost in an apparatus including a plurality of reconfigurable devices. The purpose is to provide.

According to a first aspect of the present invention, a plurality of reconfigurable devices, a redundant reconfigurable device shared by the plurality of reconfigurable devices, and a first that is one of the plurality of reconfigurable devices. When the first process is reconfigured in one reconfigurable device, the same first process is reconfigured in the redundant reconfigurable device, and the first reconfigurable device and the redundant reconfiguration are reconfigured. Comparison control for detecting a failure by comparing input / output information with a reconfigurable device and reconfiguring a process other than the first process in the reconfigurable device separate from the first reconfigurable device And a highly reliable apparatus.

  According to the second aspect of the present invention, in a data center having a group of servers having one or more reconfigurable devices as a reconfigurable node group, the first aspect of the present invention described above. In the data center, the comparison control means controls the reconfigurable device included in the reconfigurable node group.

According to a third aspect of the present invention, in the high reliability method in an apparatus including a plurality of reconfigurable devices and a redundant reconfigurable device shared by the plurality of reconfigurable devices, When the first process is reconfigured in the first reconfigurable device that is one of the reconfigurable devices, the same first process is reconfigured in the redundant reconfigurable device, and the first process is reconfigured. A failure is detected by comparing input / output information between one reconfigurable device and the redundant reconfigurable device, and the first reconfigurable device is separated from the first reconfigurable device. A highly reliable method characterized by reconfiguring processes other than the above process is provided.

According to the fourth aspect of the present invention, in the high reliability program incorporated in the high reliability apparatus, a computer can be redundantly reconfigured shared by a plurality of reconfigurable devices and the plurality of reconfigurable devices. When the first process is reconfigured for a device and a first reconfigurable device that is one of the plurality of reconfigurable devices, the same first process is reconfigured for the redundant reconfigurable device. The failure is detected by comparing the input / output information between the first reconfigurable device and the redundant reconfigurable device, and the reconfigurable device is separate from the first reconfigurable device. There is provided a highly reliable program that causes a device to function as a highly reliable device having comparison control means for reconfiguring processes other than the first process .

  ADVANTAGE OF THE INVENTION According to this invention, it becomes possible to make highly reliable the apparatus provided with two or more reconfigurable devices at low cost.

It is a figure showing the basic composition of the high reliability device concerning the embodiment of the present invention. FIG. 3 is a diagram illustrating a basic operation when assigning a process of calculation A in the high reliability device shown in FIG. 1. It is a figure showing the operation | movement condition of FIG. FIG. 3 is a diagram illustrating a basic operation in the case of assigning a process of calculation B in the high reliability device shown in FIG. 1. It is a figure showing the operation | movement condition of FIG. FIG. 2 is a diagram illustrating a basic operation in the case of assigning a process called calculation C in the high reliability device shown in FIG. 1. It is a figure showing the operation | movement condition of FIG. FIG. 3 is a diagram illustrating a basic operation when assigning a process of calculation A in the high reliability device shown in FIG. 1. It is a figure showing the operation | movement condition of FIG. FIG. 2 is a diagram illustrating a basic operation in the case of performing a test by assigning processes B and C and releasing the process A to the highly reliable apparatus illustrated in FIG. 1. It is a figure showing the operation | movement condition of FIG. FIG. 3 is a diagram illustrating a basic operation in a case where a test called calculation A is assigned and a process called calculation B is released in the high reliability device shown in FIG. 1. It is a figure showing the operation | movement condition of FIG. It is a figure explaining the structure of the data center which is an example of embodiment of this invention. It is a figure explaining the structure of the semiconductor integrated circuit which is an example of embodiment of this invention. It is a figure which shows the structure of the technique relevant to this invention.

  First, an outline of an embodiment of the present invention will be described. The embodiment of the present invention is an apparatus generally including a plurality of reconfigurable devices, a shared reconfigurable device (redundant reconfigurable device), and a comparison control unit. At the time of reconfiguration to reconfigure the process to the reconfigurable device, the comparison control unit reconfigures the same process to the shared reconfigurable device, and detects the failure by comparing the input / output information of both.

  Next, embodiments of the present invention will be described in detail with reference to the drawings.

  Referring to FIG. 1, according to a preferred embodiment of the present invention, a redundant reconfigurable device is provided in a plurality of reconfigurable devices, a comparison is performed at the time of allocation, or a test is performed at the time of release. It is.

  FIG. 1 is a diagram illustrating a configuration of a highly reliable apparatus 1000 having a reconfigurable device according to an embodiment of the present invention. Referring to FIG. 1, the highly reliable apparatus 1000 includes a plurality of reconfigurable devices 200-1 to -n, and further includes a comparison control unit 100, a redundant reconfigurable device 300, a processor 400, A main storage device 500 is included. In the figure, a reconfigurable device (first reconfigurable device) 200-1, a reconfigurable device (second reconfigurable device) 200-2, and a reconfigurable device 200-n are shown. Although illustrated, the number of reconfigurable devices 200 of the present embodiment is not limited to three, and it is possible to have any number of reconfigurable devices.

  Here, the reconfigurable device 200 refers to a device whose hardware configuration can be freely changed by given configuration information, represented by an FPGA (Field Programmable Gate Array).

  The redundant reconfigurable device 300 has the same configuration as the reconfigurable device 200. In the present embodiment, a case where the number of redundantly reconfigurable devices is one will be described, but this is merely an example. Of course, the redundant number may be increased as long as the cost is commensurate.

  The comparison control unit 100 is a part that reconfigures the reconfigurable device 200 and the redundant reconfigurable device 300. In addition, the comparison control unit 100 compares processing results by controlling input / output of information necessary for processing on the reconfigurable device 300. Further, the comparison control unit 100 may have a function of testing a reconfigurable device.

  The processor 400 is an arithmetic processing unit represented by a CPU (Central Processing Unit). When the present embodiment is realized as a combination of hardware and software, the processor 400 accesses the main storage device 500 and executes instructions stored in the main storage device 500 to thereby execute the comparison control unit 100. Is realized. Note that the configuration example of FIG. 1 is merely an example. In addition, as will be described later with reference to FIGS. 14 and 15, for example, the present embodiment is a system using a large number of devices, or a part thereof is a chip such as a SoC (System-on-a-chip). It can also be realized.

  Next, an example of the operation of the high reliability device 1000 according to the present embodiment will be described with reference to FIGS. 2 to 13, the reconfigurable device 200-n, the processor 400, and the main storage device 500 are omitted for convenience of explanation.

  2 and 3 are diagrams illustrating an example of the operation of the high reliability apparatus 1000 illustrated in FIG. In FIG. 2, the sign consisting of S and a number beside the arrow represents the step number. In this example, a case will be described in which the process of calculation A (first process) is assigned to the reconfigurable device (first reconfigurable device) 200-1.

  First, the comparison control unit 100 reconfigures the process of calculation A to the reconfigurable device 200-1. At that time, the same processing is also reconfigured in the redundant reconfigurable device 300 (step S11).

  Then, the comparison control unit 100 performs control so that the input / output information necessary for the calculation A is connected to the reconfigurable device 200-1 and the redundant reconfigurable device 300 (step S12).

  Thereby, if the outputs of the two reconfigurable devices 200-1 and 300 are different, a failure is detected.

  4 and 5 are diagrams illustrating examples of operations and status changes of the high reliability device 1000 illustrated in FIG. In FIG. 4, the sign consisting of S and a number beside the arrow represents a step number. In this example, a case where the process of calculation B (second process) is assigned to the reconfigurable device (second reconfigurable device) 200-2 will be described.

  First, the comparison control unit 100 reconfigures the process of calculation B to the reconfigurable device 200-2. At that time, the same processing is also reconfigured in the redundant reconfigurable device 300 (step S21).

  Then, the comparison control unit 100 performs control so that the input / output information necessary for the calculation B is connected to the reconfigurable device 200-2 and the redundant reconfigurable device 300 (step S22).

  As a result, if the outputs of the two reconfigurable devices 200-2 and 300 are different, a failure is detected.

  That is, through the operations of FIGS. 2 and 4, it is possible to check whether or not there is a failure in the two reconfigurable devices 200-1 and 200-2 by using only one redundant reconfigurable device. The reliability of the reconfigurable device can be improved.

  6 and 7 are diagrams illustrating examples of operations and status changes of the high reliability device 1000 illustrated in FIG. In FIG. 6, a sign consisting of S and a number beside the arrow represents a step number. In this example, a case will be described in which the process of calculation C (third process) is assigned to the reconfigurable device 200-1.

  First, the comparison control unit 100 reconfigures the process of calculation C to the reconfigurable device 200-1. At that time, the same processing is also reconfigured in the redundant reconfigurable device 300 (step S31).

  Then, the comparison control unit 100 performs control so that the input / output information necessary for the calculation C is connected to the reconfigurable device 200-1 and the redundant reconfigurable device 300 (step S32).

  Thereby, if the outputs of the two reconfigurable devices 200-1 and 300 are different, a failure is detected.

  That is, FIGS. 6 and 7 show that a failure can be detected each time a reconfigurable device is reconfigured.

  Although the case of allocation has been described above, the case of release will be described below.

  8 and 9 are diagrams illustrating examples of operations and status changes of the high reliability device 1000 illustrated in FIG. In FIG. 8, the sign consisting of S and a number beside the arrow represents a step number. In this example, a case will be described in which the process of calculation A (first process) is assigned to the reconfigurable device 200-1.

  The comparison control unit 100 reconfigures the process of calculation A to the reconfigurable device 200-1 (step S41).

  10 and 11 are diagrams illustrating examples of operations and status changes of the high reliability device 1000 illustrated in FIG. In FIG. 10, the sign consisting of S and a number beside the arrow represents a step number. In this example, a case will be described in which a process called calculation A (first process) is released and a process called calculation B (second process) and a process called calculation C (third process) are allocated.

  First, the comparison control unit 100 reconfigures the process of calculation B to the reconfigurable device 200-2 (step S51).

  Similarly, the comparison control unit 100 reconfigures the process of calculation C to the redundant reconfigurable device 300 (step S52).

  Then, the comparison control unit 100 releases the process of calculation A from the reconfigurable device 200-1 and performs a test of the reconfigurable device 200-1. Thereby, failure detection of the reconfigurable device 200-1 is possible (step S53).

  Originally, the process of calculation C must be assigned to the reconfigurable device 200-1, but since there is a redundant reconfigurable device 300, the calculation C is assigned to the calculation C when the process of calculation A is released. It is possible to test the reconfigurable device 200-1.

  12 and 13 are diagrams illustrating examples of operations and status changes of the high reliability device 1000 illustrated in FIG. In FIG. 12, the sign consisting of S and a number beside the arrow represents a step number. In this example, a case will be described in which a process called calculation B (second process) is released and a process called calculation A (first process) is assigned.

  First, the comparison control unit 100 reconfigures the process of calculation A to the reconfigurable device 200-1 that has been tested (step S61).

  Then, the comparison control unit 100 releases the process of calculation B from the reconfigurable device 200-2 and performs a test of the reconfigurable device 200-2. Thereby, the failure detection of the reconfigurable device 200-2 is possible (step S62).

  In other words, when the reconfigurable device is tested at the time of release, all other reconfigurable devices are used, and if a new process allocation is requested during the test, the test ends. The new process cannot be assigned. That is, since there is a reconfigurable device that has not been processed, the peak performance disclosed to the user cannot be realized. However, as in the examples of FIGS. 10, 11, 12, and 13, even in the above situation, by utilizing redundant reconfigurable devices, a state in which all reconfigurable devices are used is practically used. Can be reproduced. Therefore, it is possible to perform the test while providing the peak performance disclosed to the user.

  Moreover, in the above embodiment, although the fault detection by the test by duplication was demonstrated, this invention is applicable also to techniques, such as fault masking by triple of a reconfigurable device, for example.

  FIG. 14 is a diagram illustrating the configuration of a system to which the present invention can be applied. The data center in FIG. 14 includes a calculation node group 2000 and a communication node group 3000, and further includes a reconfigurable node group 4000.

  The computation node group 2000 includes servers 2000-1 to 2000-n. The servers 2000-1 to 2000-n perform processing and response to the request.

  The communication node group 3000 includes routers 3000-1 to 3000-n. The routers 3000-1 to 3000-n are used for connection between the servers 2000-1 to 2000-n and the reconfigurable node group 4000.

  The reconfigurable node group 4000 includes a reconfigurable node control unit 4100 and reconfigurable servers 4200-1 to 4200-n. This reconfigurable node group 4000 is shared by the calculation node group 2000 and the communication node group 3000.

  The reconfigurable servers 4200-1 to 4200-n have one or more reconfigurable devices, and are mainly used for accelerating processes from the computing node group 2000 and the communication node group 3000.

  The reconfigurable node control processing unit 4100 corresponds to the comparison control unit 100. Further, in response to requests from the computing node group 2000 and the communication node group 3000, reconfigurable devices are allocated and released in units of reconfigurable servers and reconfigurable devices. Here, a plurality of reconfigurable node control units 4100 may be included in the reconfigurable node group 4000.

  FIG. 15 is a diagram showing a configuration of a reconfigurable semiconductor integrated circuit 4300 used in the reconfigurable servers 4200-1 to 4200-n. The reconfigurable semiconductor integrated circuit 4300 can be realized as a SoC (System-on-a-chip), for example.

  Referring to FIG. 15, the reconfigurable semiconductor integrated circuit 4300 includes reconfigurable devices 4310-1 to 4310-n. In other words, the reconfigurable semiconductor integrated circuit 4300 is a semiconductor integrated circuit configured only from a reconfigurable device.

  Here, the redundant reconfigurable devices 4310-1 to 4310-n may be provided in the reconfigurable semiconductor integrated circuit 4300, or the reconfigurable semiconductor integrated circuit 4300 may be reconfigured as an extra server 4200. -1 to 4200-n may be included. Of course, there may be multiple redundant reconfigurable devices.

  In the configuration having redundant reconfigurable devices, this embodiment can be applied between reconfigurable devices in a semiconductor integrated circuit or between reconfigurable devices between semiconductor integrated circuits. .

  The reconfigurable servers 4200-1 to 4200-n use one or a plurality of reconfigurable equipment semiconductor integrated circuits 4300.

  In each of the above embodiments, an apparatus and a method having a plurality of reconfigurable devices have been described as examples. However, the present embodiment is not limited to such an apparatus and method, and can be applied to any apparatus and method. is there.

  Although the present invention has been described with reference to the above embodiment, the present invention is not limited only to the configuration and operation of the above embodiment, and can be made by those skilled in the art within the scope of the present invention. Of course, it includes various possible variations and modifications.

  Furthermore, although the high reliability apparatus which is the embodiment of the present invention can be realized by hardware, the computer reads a program for causing the computer to function as the high reliability apparatus from a computer-readable recording medium. It can also be realized by executing.

  Further, the high reliability method according to the embodiment of the present invention can be realized by hardware, but the computer reads a program for causing the computer to execute the method from a computer-readable recording medium and executes the program. Can also be realized.

  A part or all of the above embodiment can be described as in the following supplementary notes, but is not limited to the following.

  (Appendix 1) A plurality of reconfigurable devices, a redundant reconfigurable device shared by the plurality of reconfigurable devices, and a first reconfigurable device that is one of the plurality of reconfigurable devices When reconfiguring the first process, the same first process is reconfigured for the redundant reconfigurable device, and input / output information between the first reconfigurable device and the redundant reconfigurable device And a comparison control unit for detecting a failure by comparing the two.

  (Supplementary note 2) In the high reliability apparatus according to supplementary note 1, the comparison control unit may reconfigure the first process on the first reconfigurable device, and then reconfigure the first reconfigurable device. Reconfigures the second process to a second reconfigurable device that is a separate reconfigurable device and reconfigures the second process to the redundant reconfigurable device, and reconfigures the second process A high-reliability apparatus characterized by detecting a failure by comparing input / output information between a possible device and the redundant reconfigurable device.

  (Supplementary Note 3) In the high reliability apparatus according to Supplementary Note 1 or 2, the comparison control unit may change the reconstructed process to one of the plurality of reconfigurable devices and the redundant reconfigurable device. A highly reliable apparatus that is modified and reconstructed into one of the plurality of reconfigurable devices and the redundant reconfigurable device again.

  (Appendix 4) In the high reliability device according to any one of appendices 1 to 3, the comparison control unit reconfigures the first process in the first reconfigurable device, and After restructuring a second process on a second reconfigurable device, which is the reconfigurable device separate from the reconfigurable device, and reconfiguring a third process on the redundant reconfigurable device; A high-reliability apparatus that releases a first reconfigurable device and tests the first reconfigurable device.

  (Supplementary Note 5) In the high reliability apparatus according to Supplementary Note 4, after the test of the first reconfigurable device is completed, the comparison control unit newly assigns a process to the first reconfigurable device, Thereafter, the high-reliability apparatus is characterized in that the second reconfigurable device is tested.

  (Supplementary note 6) The high reliability apparatus according to any one of supplementary notes 1 to 5, in a data center having a group of servers having one or more reconfigurable devices as a reconfigurable node group The data center according to claim 1, wherein the comparison control means controls the reconfigurable device of the reconfigurable node group.

  (Supplementary note 7) In a highly reliable method in an apparatus comprising a plurality of reconfigurable devices and a redundant reconfigurable device shared by the plurality of reconfigurable devices, one of the plurality of reconfigurable devices When the first process is reconfigured in the first reconfigurable device, the same first process is reconfigured in the redundant reconfigurable device, and the first reconfigurable device A high reliability method characterized by detecting a failure by comparing input / output information with the redundant reconfigurable device.

  (Supplementary note 8) In the high reliability method according to supplementary note 7, after reconfiguring the first process in the first reconfigurable device, the reconfiguration separate from the first reconfigurable device The second process is reconfigured in the second reconfigurable device that is a reconfigurable device, and the second process is reconfigured in the redundant reconfigurable device, and the second reconfigurable device and the redundant reconfigurable device are reconfigured. A high reliability method characterized by detecting a failure by comparing input / output information with a configurable device.

  (Supplementary note 9) In the high reliability method according to supplementary note 7 or 8, the reconstructed process is changed to a different process in one of the plurality of reconfigurable devices and the redundant reconfigurable device, and the process is performed again. A highly reliable method comprising reconstructing one of a plurality of reconfigurable devices and the redundant reconfigurable device.

  (Supplementary note 10) In the high reliability method according to any one of supplementary notes 7 to 9, the first process is reconfigured in the first reconfigurable device, and the first reconfigurable device is Reconfiguring a second process on a second reconfigurable device that is a separate reconfigurable device and reconfiguring a third process on the redundant reconfigurable device, then reconfiguring the first process A high reliability method characterized by releasing a device and testing the first reconfigurable device.

  (Supplementary note 11) In the high reliability method according to supplementary note 10, after the test of the first reconfigurable device is finished, a process is newly assigned to the first reconfigurable device, and then the second reconfigurable device A highly reliable method characterized by testing a reconfigurable device.

  (Supplementary note 12) In a high reliability program incorporated in a high reliability apparatus, a computer is configured to include a plurality of reconfigurable devices, a redundant reconfigurable device shared by the plurality of reconfigurable devices, and the plurality of reconfigurable devices. When the first process is reconfigured in a first reconfigurable device that is one of the configurable devices, the same first process is reconfigured in the redundant reconfigurable device, and the first process is reconfigured. A high-reliability program that functions as a high-reliability device having a comparison control unit that detects a failure by comparing input / output information between a reconfigurable device and the redundant reconfigurable device.

  (Supplementary note 13) In the high reliability program according to supplementary note 12, the comparison control unit reconfigures the first process in the first reconfigurable device, and then compares the first process with the first reconfigurable device. Reconfigures the second process to a second reconfigurable device that is a separate reconfigurable device and reconfigures the second process to the redundant reconfigurable device, and reconfigures the second process A high-reliability program characterized by detecting a failure by comparing input / output information between a readable device and the redundantly reconfigurable device.

  (Supplementary Note 14) In the high reliability program according to Supplementary Note 12 or 13, the comparison control unit may change the reconstructed process to one of the plurality of reconfigurable devices and the redundant reconfigurable device. A highly reliable program that is modified and reconstructed into one of the plurality of reconfigurable devices and the redundant reconfigurable device again.

  (Supplementary note 15) In the high reliability program according to any one of Supplementary notes 12 to 14, the comparison control unit reconfigures the first process in the first reconfigurable device, and After restructuring a second process on a second reconfigurable device, which is the reconfigurable device separate from the reconfigurable device, and reconfiguring a third process on the redundant reconfigurable device; A high-reliability program characterized by releasing a first reconfigurable device and testing the first reconfigurable device.

  (Supplementary note 16) In the high reliability program according to supplementary note 15, the comparison control unit assigns a new process to the first reconfigurable device after the test of the first reconfigurable device is completed, A highly reliable program for performing a test of the second reconfigurable device thereafter.

  This application claims the priority on the basis of Japanese application Japanese Patent Application No. 2009-217051 for which it applied on September 18, 2009, and takes in those the indications of all here.

100 Comparison control unit 200-1, 200-2, 200-n Reconfigurable device 300 Redundant reconfigurable device 1000 High-reliability apparatus 2000 Compute node group 2000-1, 2000-2, 2000-n Server 3000 Communication node group 3000-1, 3000-2, 3000-3, 3000-n router 4000 reconfigurable node group 4100 reconfigurable node control unit 4200-1, 4200-2, 4200-n reconfigurable server 4300 reconfigurable semiconductor integration Circuits 4310-1, 4310-2, 4310-n Reconfigurable device 5100, 5200 Processor 5300 Comparison unit

Claims (16)

Multiple reconfigurable devices;
A redundant reconfigurable device shared by the plurality of reconfigurable devices;
When reconfiguring a first process on a first reconfigurable device that is one of the plurality of reconfigurable devices, the same first process is reconfigured on the redundant reconfigurable device, A failure is detected by comparing input / output information between the first reconfigurable device and the redundant reconfigurable device, and the reconfigurable device is separate from the first reconfigurable device. A high-reliability apparatus comprising comparison control means for reconfiguring processes other than the first process . The high reliability device according to claim 1,
The comparison control means reconfigures the first process in the first reconfigurable device, and then reconfigures the second reconfiguration that is the reconfigurable device separate from the first reconfigurable device. The second process is reconstructed in the reconfigurable device, the second process is reconfigured in the redundant reconfigurable device, and input / output information between the second reconfigurable device and the redundant reconfigurable device is obtained. A highly reliable device characterized by detecting a failure by comparing. The high reliability device according to claim 1 or 2,
The comparison control unit changes one of the plurality of reconfigurable devices and the redundantly reconfigurable device into a different process and re-configures one of the plurality of reconfigurable devices and the plurality of reconfigurable devices. A highly reliable apparatus characterized by reconstructing a redundantly reconfigurable device. The high reliability device according to any one of claims 1 to 3,
The comparison control means reconfigures the first process to the first reconfigurable device, and the second reconfigurable device is the reconfigurable device that is separate from the first reconfigurable device. After reconstructing the second process on the device and reconstructing the third process on the redundant reconfigurable device, release the first reconfigurable device and test the first reconfigurable device A highly reliable device characterized by The high reliability device according to claim 4,
The comparison control means assigns a new process to the first reconfigurable device after the test of the first reconfigurable device is completed, and then tests the second reconfigurable device. Highly reliable equipment characterized by In a data center having a group of servers having one or more reconfigurable devices as a group of reconfigurable nodes,
6. The data center according to claim 1, wherein the comparison control unit in the highly reliable apparatus controls the reconfigurable device included in the reconfigurable node group. . In a highly reliable method in an apparatus comprising a plurality of reconfigurable devices and a redundant reconfigurable device shared by the plurality of reconfigurable devices,
When reconfiguring a first process on a first reconfigurable device that is one of the plurality of reconfigurable devices, the same first process is reconfigured on the redundant reconfigurable device, A failure is detected by comparing input / output information between the first reconfigurable device and the redundant reconfigurable device, and the reconfigurable device is separate from the first reconfigurable device. A high reliability method characterized by reconfiguring processes other than the first process. The high reliability method according to claim 7,
After reconfiguring the first process on the first reconfigurable device, the second reconfigurable device is a second reconfigurable device that is separate from the first reconfigurable device. Reconstructing the process and reconfiguring the second process in the redundant reconfigurable device, and comparing the input / output information between the second reconfigurable device and the redundant reconfigurable device A highly reliable method characterized by detecting. In the high reliability method according to claim 7 or 8,
One of the plurality of reconfigurable devices and the redundant reconfigurable device are changed to a different process, and the one of the plurality of reconfigurable devices and the redundant reconfigurable device are changed again. A highly reliable method characterized by rebuilding. The high reliability method according to any one of claims 7 to 9,
Reconfiguring the first process on the first reconfigurable device and second process on a second reconfigurable device that is the reconfigurable device separate from the first reconfigurable device And after the third process is reconstructed in the redundant reconfigurable device, the first reconfigurable device is released and the first reconfigurable device is tested. High reliability method. The high reliability method according to claim 10,
After the test of the first reconfigurable device is completed, a new process is assigned to the first reconfigurable device, and then the second reconfigurable device is tested. Method. In the high reliability program incorporated in the high reliability device,
Computer
Multiple reconfigurable devices;
A redundant reconfigurable device shared by the plurality of reconfigurable devices;
When reconfiguring a first process on a first reconfigurable device that is one of the plurality of reconfigurable devices, the same first process is reconfigured on the redundant reconfigurable device, A failure is detected by comparing input / output information between the first reconfigurable device and the redundant reconfigurable device, and the reconfigurable device is separate from the first reconfigurable device. A high-reliability program that functions as a high-reliability device having comparison control means for reconfiguring processes other than the first process . In the high reliability program according to claim 12,
The comparison control means reconfigures the first process in the first reconfigurable device, and then reconfigures the second reconfiguration that is the reconfigurable device separate from the first reconfigurable device. The second process is reconstructed in the reconfigurable device, the second process is reconfigured in the redundant reconfigurable device, and input / output information between the second reconfigurable device and the redundant reconfigurable device is obtained. A highly reliable program characterized by detecting a failure by comparing. In the high reliability program according to claim 12 or 13,
The comparison control unit changes one of the plurality of reconfigurable devices and the redundantly reconfigurable device into a different process and re-configures one of the plurality of reconfigurable devices and the plurality of reconfigurable devices. A highly reliable program characterized by rebuilding into a redundant reconfigurable device. The high reliability program according to any one of claims 12 to 14,
The comparison control means reconfigures the first process to the first reconfigurable device, and the second reconfigurable device is the reconfigurable device that is separate from the first reconfigurable device. After reconstructing the second process on the device and reconstructing the third process on the redundant reconfigurable device, release the first reconfigurable device and test the first reconfigurable device A highly reliable program characterized by The high reliability program according to claim 15,
The comparison control means assigns a new process to the first reconfigurable device after the test of the first reconfigurable device is completed, and then tests the second reconfigurable device. High reliability program characterized by

Images