JP3848268B2 - Computer system, computer apparatus, data access method and program in computer system - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a computer system, a computer apparatus, and a data access method and program in the computer system, and more particularly, to a computer system, a computer apparatus, and a data access method and program in a computer system in which a software program is migrated between a plurality of computers. .
[0002]
[Prior art]
In a computer system composed of a plurality of computers capable of communicating with each other, there is a migration technique for transferring a software program (hereinafter referred to as a program) between computers.
[0003]
In the program migration technology, the context that is the execution state of the running program is saved, the context is transferred to another computer, and the execution state is restored in the transfer destination computer so that the program can be continuously executed. Technology. The program context generally means a register value of a central processing unit (hereinafter also referred to as a CPU), the contents of a memory area used by the program, and a communication state with the outside.
[0004]
The migration technique is used for high reliability and load distribution of a computer system composed of a plurality of computers. For example, when a computer fails, the program running on the failed computer is transferred to a normal computer in the system by the migration technique, and the execution is continued, so that the system is highly reliable. Alternatively, if the load is concentrated on a specific computer in the system, the program on the computer on which the load is concentrated is transferred to a computer with a low load in the system by migration technology, and the system is continued by executing it. Load distribution is realized.
[0005]
Various representative methods of program migration are disclosed (for example, see Non-Patent Document 1).
[0006]
When using the migration technique, before and after migration, a program executed in the migration destination computer may need to continuously access data on a storage medium included in the migration destination computer. In such a case, during program migration, in addition to the program context, the contents of the data on the storage medium are also replicated on the storage medium of the destination computer, and data access after the transfer is performed on the destination computer. By doing so, continuous data access before and after the migration becomes possible. In this way, if the content of the data on the storage medium is copied, subsequent data access can be performed at high speed.
[0007]
On the other hand, a network file system is often used in order to continuously access data on a storage medium before and after migration without copying data on the storage medium to a migration destination computer. A typical network file system is NFS (Network File System). NFS technology is disclosed in various documents (for example, see Non-Patent Document 2).
[0008]
If a network file system is used, it becomes possible to access data on a storage medium in the same manner from any computer constituting the computer system. The network file system is a technique that enables obtaining or updating desired data by communicating with a computer in which data entities are stored during data access.
[0009]
If the network file system is used, the data on the storage medium used by the program executed on the migration source computer can be used also by the migration destination computer. In this way, the data on the storage medium used by the program to be migrated is stored on the network file system, and the data that is visible to each computer is used to make the data on the storage medium at the time of migration. In many cases, the transfer is not performed.
[0010]
However, the form of a computer system using this network file system is premised on that computers constituting the system are connected by a high-speed network. This is because, when reading or writing data, because network communication for reference or update occurs with the computer where the substance of the data is stored, in an environment where low-speed network connections are mixed, This is because reading and writing data becomes very slow. In general, the data read / write performance greatly affects the execution speed of the program.
[0011]
[Non-Patent Document 1]
"Butterfield, SA and Popec, GJ", "Network Tasking in the Locus Distributed Unix System", Eunix Summer 84 ( Usenix Summer '84), pp.62-71 (1984)
[0012]
[Non-Patent Document 2]
Callaghan, Brent, NFS Illustrated, ISBN: 0-201-32570-5, Addison Wesley
[0013]
[Problems to be solved by the invention]
However, when copying data on a storage medium during program migration, if the data to be copied is huge, network communication for data replication takes time, and as a result, the time required for program transfer It will increase. As a result, since it is necessary to stop the program during the program migration, there is a problem that if the program migration takes a long time, the program stop time becomes long.
[0014]
On the other hand, when using a system that uses a network file system to share data on storage media between computers without duplicating data during program migration, every time the program of the migration destination computer accesses the data There is a problem that the speed of data access decreases because network communication occurs with a computer that stores the actual state of data.
[0015]
Therefore, an object of the present invention is to provide a computer system that can perform data access without reducing the data access speed in a system in which program migration is performed.
[0016]
[Means for Solving the Problems]
The computer system of the present invention is a computer system including a plurality of computer devices connected to each other via a communication path, and each computer device is connected via a communication path. First Other computer equipment And second other computer device A communication means connected to the storage device, a storage means for storing own computer identification information indicating the own computer device, First From another computer device, the program First Other computer devices are available And storing data to be referred to when executing the program. Data area information indicating a data area, and the data area information associated with the data area information First Indicates other computer equipment First Other computer identification information, Second other computer identification information that stores the same data as the first data region and is associated with a second data region of the second other computer device and indicates the second other computer device; , Migration means that receives the transfer of the data, and when referring to the data according to the program First Other computer identification information And the second other computer identification information In addition to the management means for storing the computer identification information in association with the data area information, First The other computer apparatus has First Data area Alternatively, the second data area of the second other computer device Get from Said Storage means for storing data.
[0017]
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention will be described below with reference to the drawings.
[0018]
First, based on FIG. 1, the structure of the computer system concerning this Embodiment is demonstrated. FIG. 1 is a configuration diagram showing a configuration of a computer system according to the present embodiment.
[0019]
In FIG. 1, reference numeral 100 denotes a computer system, which includes a computer 101, a computer 111, and a computer 121. Each computer is connected by a communication path 131. 110 is a communication path for connecting the computer 101 and the communication path 131, 120 is a communication path for connecting the computer 111 and the communication path 131, and 130 is a connection between the computer 121 and the communication path 131. This is a communication channel. Each communication path is connected via a connection device (not shown).
[0020]
The computer 101 includes a program unit 102. The program unit 102 includes a program main body 102a to be migrated, a location management unit 103, and a location management table 104 as will be described later. The computer 101 further includes an input / output unit 105, a storage medium 106 such as a hard disk, a program transmission / reception unit 107, and a communication device 109.
[0021]
The input / output unit 105 is a program for inputting / outputting data to / from the storage medium 106 in response to a command from the program unit 102. The program transmission / reception unit 107 is a program that performs migration of a program and generation of a moved or transferred program. The program transmission / reception unit 107 includes a context 108 for migration of the program unit 102 and manages the context 108. The communication device 109 performs data communication with other computers via the communication path 131.
[0022]
The computer 101 includes a location management unit 103 that manages location information of data accessed by the program main body 102 a and a location management table 104. To the unit 105. The location management unit 103 is a part of the program unit 102 and manages a location management table 104 that stores location data indicating which computer has data used by the program main body 102a. The location management table 104 is in the memory used by the program unit 102 and is included in the context 108.
[0023]
For example, when the program body 102 a writes data to a certain data area on the storage medium 106 via the input / output unit 105, the write request access data is supplied to the input / output unit 105 via the position management unit 103. In response to the write request, the input / output unit 105 writes data in the data area on the storage medium 106.
[0024]
The computer 111 includes an input / output unit 115, a storage medium 116 such as a hard disk, a program transmission / reception unit 117, and a communication device 119. Similar to the input / output unit 105, the input / output unit 115 is a program that performs data input / output with respect to the storage medium 116. Similar to the program transmission / reception unit 107, the program transmission / reception unit 117 is a program for performing migration of the program unit 102 and generating a transferred program. The program transmission / reception unit 117 manages the context of the program unit 102. Similar to the communication apparatus 109, the communication apparatus 119 performs data communication with other computers via the communication path 131.
[0025]
The computer 121 includes an input / output unit 125, a storage medium 126 such as a hard disk, a program transmission / reception unit 127, and a communication device 129. The input / output unit 125 is a program for inputting / outputting data to / from the storage medium 126, similarly to the input / output unit 105. Similarly to the program transmission / reception unit 107, the program transmission / reception unit 127 is a program that performs program migration and generation of a transferred program. The program transmission / reception unit 127 manages the context of the program unit 102. Similar to the communication device 109, the communication device 129 performs data communication with other computers via the communication path 131.
[0026]
Thus, since each computer has a program transmission / reception unit, the computer system 100 has a function of performing program migration between computers. Accordingly, in the computer system 100, the program main body 102a can operate while the program unit 102 is being transferred between the plurality of computers 101, 111, and 121.
[0027]
Each of the computers 101, 111, and 121 is preliminarily assigned an identifier that is unique computer identification information, and each computer can be specified by the identifier. Each computer stores data of a computer identifier indicating its own computer in a memory device which is data storage means (not shown). The location management unit 103 can acquire the identifier of the computer executing the program unit 102. The computer identifier may be the host name or IP address of the computer, but the type of the computer identifier is not limited as long as it is unique to each computer.
[0028]
The location management unit 103 obtains information on an area in which data in the storage medium 106 is written by monitoring the contents of access performed by the program body 102 a to the input / output unit 105. The location management unit 103 stores and manages, in the location management table 104, a combination of data area information accessed by the program main body 102a and the identifier of the computer 101 executing the program main body 102a.
[0029]
In the position management table 104, as the position data, correspondence between data area information indicating a data area accessed by the program main body 102a and the identifier of the computer storing the data in the data area is recorded. FIG. 2 is a diagram illustrating an example of the location management table 104. In the location management table 104, for each data area information indicating a data area, an identifier of a computer indicating a specific computer having the data area is recorded in an entry corresponding to the data area. In this example, the data of the area A represents that the computer 111 has. Similarly, the data of the area B is included in the computer 111 and the computer 121, and the data of the area C is included in the computer 101. In the location management table 104 of the computer 101, the presence of the identifiers of the computers 111 and 121 indicates that the program main body 102a has operated on the computers 111 and 121 in the past and has updated the data. The table update method will be described later. The number of computer identifiers that can be recorded in the entry corresponding to each data area of the location management table 104, that is, the data registration area is not defined, and any number may be used as long as it is one or more. You may make it provide restrictions, such as leaving.
[0030]
The location management table 104 managed by the location management unit 103 is transferred to another computer as a part of the context of the program unit 102 when the program unit 102 is transferred to another computer, as will be described later.
[0031]
Although it is conceivable to use Ethernet (R), a wireless LAN, or the like as the communication paths 110, 120, 130, and 131, any type is acceptable as long as they can communicate with each other. Information on the communication speed of each communication path is held in each computer as tabular data as shown in FIG. 3, for example. FIG. 3 is a diagram illustrating an example of a speed management table that records the speed of each communication path. In this example, the speeds of the communication paths 131, 120, and 130 are 100 Mbps high-speed communication paths, and the speed of the communication path 110 is 10 Mbps low-speed communication paths. Each computer refers to this speed management table, so that the communication speed between the computer 101 and the computer 111 and between the computer 101 and the computer 121 is 10 Mbps, and the communication speed between the computer 111 and the computer 121 is It can be determined that the speed is 100 Mbps. The management method of this table is not particularly specified as long as each computer can refer to this table.
[0032]
In the present embodiment, the program migration method is not particularly limited. The migration process can be realized by transferring the CPU register value that is the execution context of the program and the contents of the memory used by the program between computers using the same method as the conventional migration, so the contents of the migration process Description of is omitted. In addition, it is possible to migrate programs between a plurality of computers 101, 111, 121 by using the program transmission / reception unit of computers 111, 121. However, since migration processing can also be performed in the same manner as before, Description is omitted.
[0033]
First, according to the present embodiment, when the program is transferred in the computer system 100 of FIG. 1, how the data in the data area accessed by the program is distributed among a plurality of computers and duplicated. It will be described using a specific example how it is stored.
[0034]
FIGS. 4 to 7 are diagrams for explaining an example of how data is replicated between computers when a program is transferred.
[0035]
As shown in FIG. 4, the program unit 102 is operating on the computer 101 in the initial state, and the data in the data areas A, B, and C are stored in the storage medium 106 provided in the computer 101, and the program unit 102 Is accessing data areas A, B, and C in the storage medium 106. Here, in the location management table 104 included in the program unit 102, it is assumed that the computer identifiers corresponding to the data areas A, B, and C are only “101”, respectively.
[0036]
Assume that the program unit 102 is transferred and migrated to the computer 111 from the initial state of FIG. 4, and the program unit 102 is continuously executed in the migration destination computer 111. When the program unit 102 operating on the computer 111 accesses to write or update the data in the data area A during the execution, the update data is transferred to the recording medium 106 as shown in FIG. Area) data is stored or written in the recording medium 116. Therefore, the program unit 102 operating on the computer 111 accesses the recording medium 116 when accessing the data area A thereafter. At this time, as shown in FIG. 5, the data in the data area A exists only in the recording medium 116 of the computer 111, and the data in the data area A recorded in the recording medium 106 of the migration source computer 101 is invalidated. The
[0037]
Therefore, the data update operation by the migrated program unit 102 is written in a storage medium included in the computer on which the program unit 102 operates, and therefore does not involve communication processing via a communication path, and thus can be performed at high speed.
[0038]
Then, if the migrated program unit 102 of the computer 111 accesses the data area B by referring to the data, that is, simply reading the data area B without updating it, the computer 101 (reference data holding computer) stores the data in the recording medium 106. Data in the data area B (reference data area) is acquired and copied to the recording medium 116 of the computer 111 as shown in FIG. Therefore, after that, when the program unit 102 operating on the computer 111 refers to the data in the data area B, it accesses the recording medium 116 and does not perform communication processing via the communication path. It can be done at high speed.
[0039]
Next, it is assumed that the program unit 102 of the computer 111 is transferred to the computer 121. Then, when the program unit 102 of the computer 121 refers to the data in the data area B, similarly to FIG. 6, the computer 121 starts the data area B (reference data area) from the computer 111 faster than the communication speed with the computer 101. Data is acquired and copied to the recording medium 126 of the computer 121 as reference data area data as shown in FIG. Both the computer 101 and the computer 111 are reference data holding computers having data in the data area B, but the communication speed of the communication path 110 is 10 Mbps and the communication speed of the communication path 120 is 100 Mbps. This is because data transfer from 111 is more efficient than data transfer from the computer 101. Thereafter, when the program unit 102 of the computer 121 refers to the data in the data area B, the recording medium 126 is accessed.
[0040]
As described above, when the data in the data area on the storage medium accessed by the program is updated by the program operating on the migration destination computer, the data in the update data area is stored in the storage medium included in the migration destination computer. The data in the corresponding data area of the migration source computer is invalidated. In addition, when data in a data area on a storage medium accessed by a program is referenced by a program operating on a migration destination computer, the data in the reference data area is transferred to the migration destination computer and duplicated data is stored. Generated.
[0041]
By performing such an operation, the data in the data area on the storage medium accessed by the migrated program is distributed and stored in multiple on a plurality of computers. Further, as will be described later, in the case of the data reference operation by the program, since the data in the reference data area is stored in a plurality of computers (reference data holding computers), from among the plurality of reference data holding computers, By transferring data in the reference data area from a computer that can transfer data at the highest speed, efficient data access can be performed.
[0042]
Next, as shown in FIG. 4 to FIG. 7 described above, the data update operation and the data reference operation are performed in the process of transferring the program on the computer system as an example. Will be described. FIG. 8 is a flowchart illustrating an example of a processing flow of the position management unit 103 in the case of data reference. FIG. 9 is a flowchart illustrating an example of a processing flow of the location management unit 103 in the case of data update. FIG. 10 is a diagram for explaining an example of how the contents of the location management table change when data is updated and referenced while the program is migrated.
[0043]
In the state of FIG. 4, in the location management table of FIG. 2, the computer identifiers of the areas A, B, and C are all “101”, and there is no computer identifier other than “101”. This is because the program unit 102 performs data access to the areas A, B, and C of the recording medium 106 before the program unit 102 is transferred to the computer 111.
[0044]
First, the flow of processing of the location management unit 103 when the program unit 102 operating on the computer 101 refers to and updates data in the data area on the storage medium 106 will be described. First, data reference operation processing will be described with reference to FIG.
[0045]
When the program main body 102a operating on the computer 101 issues a reference request for data in the data area B to the position management unit 103, the position management unit 103 receives the reference request (step 401). Next, based on the location management table 104, the location management unit 103 determines whether or not the data area referred to in the reference request (reference data area) is in its own computer (step 402). In other words, the location management unit 103 checks whether or not the identifier of the computer 101 exists in the entry corresponding to the data area B of the location management table 104. In the case of FIG. 4, since the identifier of the computer 101 is in the entry corresponding to the data area B of the location management table 104, that is, the computer 101 is a computer having data in the data area B (reference data holding computer). The position management unit 103 reads the data from its own storage medium 106 via the input / output unit 105 (step 403). Then, the location management unit 103 supplies the data received from the input / output unit 105 to the program body 102a (step 408).
[0046]
Next, data write operation, that is, update operation processing will be described with reference to FIG. When the program unit 102 operating on the computer 101 issues a write request, that is, an update request together with the write data, the location management unit 103 receives the update request (step 501). Based on the position management table 104, the position management unit 103 determines whether the data area (update data area) requested to be updated belongs to a computer other than itself (step 502). For example, when the program main body 102 a updates data in the data area A, the location management unit 103 checks whether an identifier of a computer other than the computer 101 exists in the entry corresponding to the area A in the location management table 104.
[0047]
In the case of FIG. 4, since the identifier of the computer other than the computer 101 does not exist in the entry of the data area A, the process proceeds to step 505, and the write data is written as the data area A in its own storage medium 106 ( Step 505).
[0048]
In the case of FIG. 4, since the computer identifiers corresponding to the respective data areas in the location management table 104 are all “101”, in the update operation for the data area whose data of the computer identifier is “101”, the processing is as follows. In step 502, NO is determined, and the position management unit 103 transfers an update request to the input / output unit 105, and the input / output unit 105 writes the write data of the data area to be updated to the storage medium 106 (step 505). ).
[0049]
The above is the processing flow of the location management unit 103 when the program unit 102 operating on the computer 101 refers to and updates data in the above-described initial state. Note that the order of processing in steps 502 to 504 and step 505 is not limited, and either may be performed first or simultaneously.
[0050]
When the program unit 102 operates on the computer 101 and accesses the data in the storage medium 106 of the computer 101, the state of step 611 in FIG. 10 is in effect, and the content of the location management table 104 is 601 at that time. This is the state shown.
[0051]
Next, it is assumed that the program unit 102 is migrated from the computer 101 to the computer 111 as shown in Step 612 of FIG. At this time, since the program unit 102 is transferred to the computer 111, the position management unit 103 and the position management table 104 are also transferred to the computer 111 together. The transferred program unit 102 is restored by the computer 111 and re-executed.
[0052]
When the program unit 102 is re-executed after the migration of the program 102, the program body 102a performs data access to the input / output unit via the location management unit 103. At this time, the location management unit 103 8 and FIG. 9 are executed.
[0053]
First, the processing flow of the location management unit 103 when the program main body 102a operating on the computer 111 updates data in the data area A will be described. In FIG. 9, when the program body 102a issues a data write request, that is, an update request together with the write data, the location management unit 103 receives the update request (step 501). Based on the position management table 104, the location management unit 103 determines whether the data area (update data area) for which update has been requested belongs to a computer other than itself, that is, other than the recording medium 116 of the migration destination computer 111. Is determined (step 502).
[0054]
If the program unit 102 is transferred from the computer 101 and has not yet accessed the data in each data area, the computer identifier corresponding to each data area in the location management table 104 is all “101”. Yes. Therefore, YES is obtained in step 502, and the computer identifier data “101”, which is a computer identifier other than itself, corresponding to the data area A is deleted (step 503), and the data on the storage medium 106 is invalidated. Next, in the location management table 104, the computer identifier “111” of its own computer is written as the computer identifier corresponding to the data area A for which an update request has been made (step 504). Then, according to the update request, the update request is transferred to the input / output unit 115, and the input / output unit 115 writes the write data in the data area A into its own storage medium 115 (step 505).
[0055]
Next, a flow of processing of the location management unit 103 when the program main body 102a operating on the computer 111 refers to data in the data area B will be described.
[0056]
When the program body 102a issues a reference request for data in the data area B to the position management unit 103, the position management unit 103 receives the reference request (step 401). The position management unit 103 checks whether or not the identifier of the computer 111 exists in the entry corresponding to the data area B (reference data area) of the position management table 104 (step 402). For the data area B related to the reference request, since the identifier of the computer 111 that is its own computer does not yet exist in the location management table 104, the identifier of the computer 101 exists instead. , Go to Step 404.
[0057]
In step 404, the identifier of the computer existing in the entry of the data area B which is the reference data area is checked, and if there are a plurality of identifiers, the computer that can transfer data to the computer 111 at the highest speed is selected from among the identifiers. The reference request is output to the computer. Here, since the identifier of only the computer 101 exists in the entry of the data area B of the location management table 104, the computer 101 (reference data holding computer) is selected because there is only transfer of data from the computer 101. . The location management unit 103 reads data in the data area B from the storage medium 106 through the communication device 119, the communication paths 120, 131, and 110, the communication device 109, and the input / output unit 105 (step 405). The location management unit 103 writes the acquired data area B data to the storage medium 116 via the input / output unit 115 (step 406). The location management unit 103 adds the identifier of the computer 111, which is its own computer identifier, to the entry corresponding to the data area B of the location management table 104 (step 407). Then, the data acquired in step 405 is supplied to the program body 102a (step 408).
[0058]
Therefore, after the program unit 102 is transferred from the computer 101 to the computer 111, when the migration destination computer 111 updates the data, the updated data is stored in the storage medium 116 of the computer 111 and simultaneously stored in the computer 101. The data on the medium 106 is invalidated, and when the migration destination computer 111 references the data, a copy of the referenced data is stored in the storage medium 116 (see FIG. 6).
[0059]
As described above, in the migration destination computer 111, when the program main body 102a refers to the data on the storage medium 116 via the input / output unit 115, the location management unit 103 monitors the access contents to the input / output unit 115. Thus, the computer in which the data of the referenced data area is stored is searched using the location management table 104 managed by the location management unit 103. Next, a computer that can transfer data to the computer 111 that is currently executing the program at the highest speed is selected from a list of computers that hold data found by the search (reference data holding computer), and the computer Read data from the storage medium.
[0060]
Further, when the program main body 102a acquires the data from another computer in order to refer to the data in the data area, the data is also copied to the storage medium 116 included in the computer 111 executing the program, and the position management unit The identifier of the computer 111 is added to the entry corresponding to the referenced data area in the location management table 104 managed by the node 103. As a result, data is distributed and stored in multiple on a plurality of computers constituting the system, and when a reference to this data area occurs thereafter, the storage medium of the computer 111 can also be used as a reference destination.
[0061]
As described above, when the program body 102a operating on the computer 111 updates the data in the data area A and further refers to the data in the data area B (step 613), the contents of the location management table 104 are indicated by 602. It becomes a state.
[0062]
Furthermore, it is assumed that the program unit 102 is transferred from the computer 111 to the computer 121 (step 614). Since the program unit 102 is transferred to the computer 121, the position management unit 103 and the position management table 104 are also transferred to the computer 121 together. The transferred program unit 102 is restored in the computer 121 and re-executed.
[0063]
When the program unit 102 is re-executed after the migration of the program unit 102, the program body 102 a performs data access to the input / output unit 125 via the location management unit 103 in the computer 121.
[0064]
First, the processing flow of the location management unit 103 when the program main body 102a operating on the computer 121 refers to the data in the area B will be described. When the program main body 102 a refers to data via the position management unit 103, the program main body 102 a issues a data reference request for data in the data area B to the position management unit 103. The location management unit 103 receives the reference request (step 401). The location management unit 103 checks whether the identifier of the computer 121 exists in the entry corresponding to the data area B of the location management table 104 (step 402). At this time, since the identifier of the computer 121 does not yet exist in the entry of the data area B of the location management table 104, NO is determined in step 402, and the computer that can transfer data to the computer 121 at the highest speed (reference) (Data holding computer) is extracted based on the data in the speed management table of FIG. 3, and a reference request is transmitted to the computer having the fastest transfer speed (step 404).
[0065]
According to the location management table 104, the computer identifier corresponding to the data area B is the identifier of the computer 101 and the computer 111 (both are reference data holding computers). You have to be forwarded. Therefore, the position management unit 103 determines which of the computer 101 and the computer 111 can receive the data earlier based on the speed data of each communication path in FIG. For example, since the computer 121 and the computer 101 pass through the communication paths 130, 131, and 110, the speed of the slowest communication path 110 is 10 Mbps. On the other hand, since the computer 121 and the computer 111 pass through the communication paths 130, 131, and 120, all have the same communication speed, and therefore 100 Mbps. Accordingly, the communication path 110 to which the computer 101 is connected has a communication speed slower than that of the communication paths 110, 130, and 131, and it is faster to transfer data from the computer 111, so the location management unit 103 selects the computer 111. The location management unit 103 transmits a reference request to the computer 111 with the fastest transfer speed in order to transfer data in the data area B from the computer 111 (step 404).
[0066]
The location management unit 103 reads the data in the data area B from the storage medium 116 through the communication device 129, the communication paths 130, 131, 120, the communication device 119, and the input / output unit 115 (step 405). The location management unit 103 writes the data of the data area B read and acquired through the input / output unit 125 to the storage medium 126 (step 406). The location management unit 103 adds the identifier of the computer 121 to the entry corresponding to the data area B of the location management table 104 (step 407). The data acquired in step 405 is output to the program body 102a (step 408). As a result, the data on each storage medium is as shown in FIG.
[0067]
Note that there is no restriction on the order of steps 404 to 406 and step 407, which may be performed first or simultaneously.
[0068]
Further, when the program main body 102a on the computer 121 tries to update the data in the data area A, the processing of FIG. 9 is executed.
[0069]
Since the program main body 102a updates the data in the data area A in the computer 111, the computer identifier corresponding to the data area A in the location management table 104 is “111”. Therefore, YES is determined in the step 502, and the computer identifier data “111” corresponding to the data area A is deleted (step 503). Next, in the location management table 104, the computer identifier “121” of its own computer is written as a computer identifier corresponding to the data area A (update data area) for which an update request has been made (step 504). Then, according to the update request, the update request is transferred to the input / output unit 125, and the input / output unit 125 writes the write data as data in the data area A to the storage medium 126 (step 505). As a result, the data in the data area A is written to the storage medium 126 of the computer 121, and the data in the data area A on the storage medium 116 of the computer 111 is invalidated.
[0070]
As described above, when the program body 102a refers to the data in the data area B and further updates the data in the area A (step 615), the contents of the location management table 104 are in a state indicated by 603.
[0071]
Further, it is assumed that the program unit 102 has been migrated from the computer 121 to the computer 101 (step 616). If the program main body 102a on the computer 101 refers to data in the data area B, the entry corresponding to the data area B in the position management table 104 also includes the identifier of the computer 101. 103 becomes YES in step 402, and data in the data area B to be referred to is read from the storage medium 106 in step 403.
[0072]
As described above, when the program body 102a refers to the data in the data area B (step 617), the contents of the location management table 104 are in a state indicated by 604.
[0073]
Therefore, by using the above data access method while the program unit 102 migrates between the computers 101, 111, and 121, data copies are distributed and stored in multiple on a plurality of computers. As a result, when data that is multiplexed and arranged between computers in the system increases and data becomes necessary, data is transferred from the computer that can perform data communication at the highest speed among the computers that store the data. Acquire efficient reference access by acquiring. In addition, update access can be performed at high speed because it is performed on a storage medium included in the computer on which the program operates.
[0074]
As described above, when a program is transferred between computers constituting a computer system and the program updates data, the data is stored on the storage medium of the migration destination computer. Also, when the program on the migration destination computer references the data, the data is replicated on the storage medium of the migration destination computer, and when the replication is generated, the data stored in a distributed manner is stored. By selecting and referring to a computer that can transfer data at the highest speed from among a plurality of reference data holding computers, it is possible to transfer data at high speed, thereby realizing efficient reference access.
[0075]
In the above description, the case where the program body 102a refers to the data has been described. However, when the location management unit 103 refers to the data for the purpose of prefetching the data, the above-described reference processing is performed. Also good.
[0076]
Further, as described above, the program migration method in the present embodiment, like the conventional migration method, transfers the program context to another computer and restores the program execution state on the migration destination computer. The method to be used is used. In this case, any method may be used as the method for storing and transferring the context, and the method is not limited.
[0077]
Further, the migrated program may be a normal application program, an operating system program, or the like.
[0078]
Furthermore, in the above description, the data area accessed by the program has been described as being in a storage medium provided in the computer. However, a program operating on the computer may be handled as a file, so the data area is stored in the storage medium. Includes files. In this case, the data area accessed by the program is a data area obtained by dividing the file.
[0079]
Furthermore, in the present embodiment, the position management unit or the position management table has been described as a part of the program, but may not be included in the migrated program. In that case, the location management unit or location management table may be sent from the migration source computer to the migration destination computer separately from the program to be migrated, or only the location management table may be sent between computers. The location management unit may communicate with the computer in advance.
[0080]
In addition, when referring to data, the timing at which the migration destination computer acquires the data from another computer is performed when the program refers to the data in the above-described embodiment. May be other than when the program refers to the data. For example, it may be a so-called data prefetch cache in which the location management unit reads data from another computer prior to a reference request from a program. That is, the above-described method can be applied not only when the program refers to data but also when the data is referred to in advance for prefetching before performing the data reference operation.
[0081]
As described above, according to the present embodiment, at the time of program migration, it is not necessary to copy the data on the storage medium to the storage medium of the destination computer. It is possible to keep the program stop time short.
[0082]
Each “unit” in this specification is a conceptual one corresponding to each function of the embodiment, and does not necessarily correspond to a specific hardware or software routine on a one-to-one basis. Therefore, in the present specification, the embodiment has been described above assuming a virtual circuit block (unit) having each function of the embodiment. In addition, each step of each procedure in the present embodiment may be executed in a different order for each execution by changing the execution order and performing a plurality of steps at the same time, as long as it does not contradict its nature.
[0083]
The program for executing the operations described above is recorded or stored in whole or in part on a portable medium such as a floppy disk, CD-ROM, or a storage device such as a hard disk. Yes. The program is read by a computer, and all or part of the operation is executed. Alternatively, all or part of the program can be distributed or provided via a communication network. The user can easily realize the computer system of the present invention by downloading the program via a communication network and installing it on a computer, or installing it from a recording medium to a computer.
[0084]
The present invention is not limited to the above-described embodiments, and various changes and modifications can be made without departing from the scope of the present invention.
[0085]
【The invention's effect】
As described above, according to the present invention, it is possible to realize a computer system that can perform data access without reducing the data access speed in a system in which program migration is performed.
[Brief description of the drawings]
FIG. 1 is a configuration diagram showing a configuration of a computer system according to an embodiment of the present invention.
FIG. 2 is a diagram showing an example of a location management table according to the embodiment of the present invention.
FIG. 3 is a diagram showing an example of a speed management table according to the embodiment of the present invention.
FIG. 4 is a diagram for explaining an example of how data is replicated between computers when a program is transferred;
FIG. 5 is a diagram for explaining an example of how data is replicated between computers when a program is transferred;
FIG. 6 is a diagram for explaining an example of how data is replicated between computers when a program is transferred;
FIG. 7 is a diagram for explaining an example of how data is replicated between computers when a program is transferred;
FIG. 8 is a flowchart showing a processing flow of a location management unit in the case of data reference according to the embodiment of the present invention.
FIG. 9 is a flowchart showing an example of a processing flow of a location management unit in the case of data update according to the embodiment of the present invention.
FIG. 10 is a diagram for explaining an example of how the contents of a location management table change when data is updated and referenced while a program is migrated.
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 100 ... Computer system, 101, 111, 121 ... Computer, 110, 120, 130, 131 ... Communication path, 106, 116, 126 ... Storage medium

Claims (16)

In a computer system including a plurality of computer devices connected to each other via a communication path,
Each computer device
Communication means connected to the first other computer device and the second other computer device via a communication path;
Storage means for storing own computer identification information indicating the own computer device;
From said first other computer device, program and, said first other computer device first data area information indicating a data area for storing data to be referenced when executing chromatic above program, in the data area information a correspondence is first other computer identification information indicating the first other computer device, the second data area the first of the stores the same the data as the data area of the second other computer device having Migration means that receives the second other computer identification information associated with the second other computer identification information indicating the second other computer device;
Said first other computer identification information and the second other computer in addition to the identification information management means for storing in association with the own computer identification information to the data area information when performing reference of the data in accordance with said program When,
Storage means for storing the data acquired from the second data area in which the first other computer device said first data area or the second other computer device having the having,
A computer system comprising: The management means further deletes the first other computer identification information and the second other computer identification information and updates the own computer identification information to the data area information when updating the data according to the program. 2. The computer system according to claim 1, wherein the computer system is stored in association with each other. In a computer system including a plurality of computer devices connected to each other via a communication path,
Each computer device
A communication means connected to another computer device via a communication path;
Storage means for storing own computer identification information indicating the own computer device;
Migration means for receiving transfer of a program, data area information indicating a data area of the other computer apparatus, and other computer identification information indicating the other computer apparatus associated with the data area information from the other computer apparatus When,
A management unit that stores the computer identification information in association with the data region information in addition to the other computer identification information when referring to the data region according to the program;
Storage means for storing data acquired from the data area of the other computer device;
With
When the management unit performs the reference according to the program and there are a plurality of other computer devices associated with the data area information, a communication speed is set among the plurality of other computer devices via the communication path. A computer system characterized by acquiring data in the data area from the fastest computer device.   The management unit is further configured to delete the other computer identification information and store the own computer identification information in association with the data region information when updating the data region according to the program. Item 4. The computer system according to item 3. Communication means connected to the first other computer device and the second other computer device via a communication path;
Storage means for storing own computer identification information indicating the own computer device;
From said first other computer device, program and, said first other computer device first data area information indicating a data area for storing data to be referenced when executing chromatic above program, in the data area information a correspondence is first other computer identification information indicating the first other computer device, the second data area the first of the stores the same the data as the data area of the second other computer device having Migration means that receives the second other computer identification information associated with the second other computer identification information indicating the second other computer device;
A management means for storing the computer identification information in association with the data area information in addition to the first other computer identification information and the second other computer identification information when referring to the data according to the program; ,
Storage means for storing the data acquired from the second data area in which the first other computer device said first data area or the second other computer device having the having,
A computer apparatus comprising: The management means further deletes the first other computer identification information and the second other computer identification information and updates the own computer identification information to the data area information when updating the data according to the program. 6. The computer apparatus according to claim 5, wherein the computer apparatus stores the information in association with each other. A communication means connected to another computer device via a communication path;
Storage means for storing own computer identification information indicating the own computer device;
Migration means for receiving transfer of a program, data area information indicating a data area of the other computer apparatus, and other computer identification information indicating the other computer apparatus associated with the data area information from the other computer apparatus When,
A management unit that stores the computer identification information in association with the data region information in addition to the other computer identification information when referring to the data region according to the program;
Storage means for storing data acquired from the data area of the other computer device;
With
When the management unit performs the reference according to the program and there are a plurality of other computer devices associated with the data area information, a communication speed is set among the plurality of other computer devices via the communication path. A computer apparatus characterized in that the data area data is acquired from the fastest computer apparatus.   The management unit is further configured to delete the other computer identification information and store the own computer identification information in association with the data region information when updating the data region according to the program. Item 8. The computer device according to Item 7. A plurality of computer devices connected to each other via a communication path, each computer device being connected to a first other computer device and a second other computer device via a communication path; storage means for storing its own computer identification information indicating a first data area for storing data wherein the first program from another computer device a first other computer devices refers when executing chromatic and said program It said second other stores the same the data as the data area information of the first other computer identification information and the first data area corresponding to the data area information indicating the first other computer apparatus shown a migration means for receiving the transfer of the second other computer identification information associated with the second data area by the computer system has shown the second other computer device, the first other computer identification A management means for storing in association distribution and the second other computer identification information to the data area information, the first other computer device having a first data area or the second other computer devices a data access method in a computer system having a storage means for storing the data acquired from the second data area,
The management means associates the own computer identification information with the data area information in addition to the first other computer identification information and the second other computer identification information when referring to the data according to the program. Remember,
It said management means to store the data acquired from the first of said first data area or the second data area in which the second other computer device having the other computer device having said storage means A data access method in a computer system. The management means further deletes the first other computer identification information and the second other computer identification information and updates the own computer identification information to the data area information when updating the data according to the program. 10. A data access method in a computer system according to claim 9 , wherein the data is stored in association with each other. A plurality of computer devices connected to each other via a communication path, each computer device storing communication means connected to other computer devices via a communication channel, and storage for storing own computer identification information indicating the own computer device And migration means for receiving transfer of a program, data area information indicating a data area of the other computer apparatus, and other computer identification information indicating the other computer apparatus associated with the data area information, from the other computer apparatus A data access method in a computer system comprising: management means for storing computer identification information in association with the data area information; and storage means for storing data acquired from the data area of the other computer device,
The management means stores the computer identification information in association with the data area information in addition to the other computer identification information when referring to the data area according to the program,
When the management unit performs the reference according to the program and there are a plurality of other computer devices associated with the data area information, a communication speed is set among the plurality of other computer devices via the communication path. Get the data area data from the fastest computer device,
The data access method in a computer system , wherein the management means stores the data acquired from the data area of the other computer device in the storage means .   The management unit is further configured to delete the other computer identification information and store the own computer identification information in association with the data region information when updating the data region according to the program. Item 12. A data access method in a computer system according to Item 11. Communication means connected to the first other computer apparatus and the second other computer apparatus via a communication path, storage means for storing own computer identification information indicating the own computer apparatus, and the first other computer apparatus other computing device the associated first data area information and the data area information indicating the first data area in which the the program first other computer device stores data to be referred to in executing chromatic and said program The first other computer identification information shown and the same data as the first data area are stored, and the second other computer apparatus is associated with the second data area of the second other computer apparatus. to the second and migration means for receiving the transfer of the other computer identification information, storing said first other computer identification information a second other computer identification information the data area information associated with showing And management means, said first computer having a storage means for storing the data acquired from the second data area in which the first data area or the second other computer device having the other computer device having,
When referring to the data according to the program by the management means, the computer identification information is associated with the data area information in addition to the first other computer identification information and the second other computer identification information. A function to memorize,
By the management unit, a function of storing the data acquired from the second data area in which the first other computer device said first data area or the second with the other computer device having said storage means A program for running When the data area is updated according to the program by the management means, the first other computer identification information and the second other computer identification information are deleted, and the own computer identification information is replaced with the data area information. The program according to claim 13 , which causes the computer to execute a function of storing in association with each other. Communication means connected to another computer apparatus via a communication path, storage means for storing own computer identification information indicating the own computer apparatus, data from the other computer apparatus and data indicating a data area of the other computer apparatus Migration means for receiving transfer of area information and other computer identification information indicating the other computer apparatus associated with the data area information; management means for storing computer identification information in association with the data area information; In a computer having storage means for storing data acquired from the data area of the computer device,
A function of storing the computer identification information in association with the data area information in addition to the other computer identification information when the management means refers to the data area according to the program;
When the management unit performs the reference according to the program and there are a plurality of other computer devices associated with the data area information, a communication speed is set among the plurality of other computer devices via the communication path. A function of acquiring data in the data area from the fastest computer device ;
A function of storing the data acquired from the data area of the other computer device in the storage unit by the management unit;
For causing the computer to execute.   When the data area is updated according to the program, the management unit further executes a function of deleting the other computer identification information and storing the own computer identification information in association with the data area information in the computer. The program according to claim 15 for causing the program to occur.

Images