JP5544963B2 - Data processing apparatus, data distributed processing system, data processing program, and data processing method - Google Patents

Description

  The present invention relates to a data processing device, a data distribution processing system, a data processing program, and a data processing method.

  Patent Document 1 discloses a method for searching the contents of a distributed directory of a system composed of a plurality of nodes arranged in a ring topology with a determined direction. In this method, a storage node of content is specified by a distributed hash function. Also, the content is duplicated at the front node for backup.

  Patent Document 2 discloses a plurality of peers connected to a network. Each peer generates a logical identifier by converting multi-attribute data into a one-dimensional value, and identifies and identifies other peers that manage the generated logical identifier based on the assignment of the logical identifier to each peer. Register data with other peers.

  Patent Document 3 discloses a technique for registering data storage location information in a node having a hash value close to a hash value generated from data.

  Non-Patent Document 1 discloses a system composed of a plurality of nodes assigned keys. The system places the object at a node where at least the first i bits of the object identifier match the key. The bit number i is calculated based on the popularity of the object. In the system, the search for an object started by each node is completed at most i hops.

  Non-Patent Document 2 discloses a data search protocol (Chord) in a system composed of a plurality of nodes that store key-corresponding data assigned to its own device.

Special table 2009-543237 JP 2008-234563 A Japanese Patent No. 4306740

V. Ramasubramanian and E. G. Sirer, “Beehive: Exploiting powerlaw query distributions for O (1) lookup performance in peer to peer overlays,” in Proceedings of the 1st USENIX Symposium on Networked Systems Design and Implementation, 2004 Ion Stoica, Robert Morris, David Karger, M. Frans Kaashoek, Hari Balakrishnan, Chord: A scalable peer-to-peer lookup service for internet applications, Proceedings of the 2001 conference on Applications, technologies, architectures, and protocols for computer communications, p.149-160, August 2001, San Diego, California, United States.

  A system composed of a plurality of nodes cannot have an appropriate multiplicity of replicated data / substitutable devices.

  The techniques described in Patent Document 1 and Non-Patent Document 1 cannot control the multiplicity of replicated data. Techniques described in other documents do not allow retention of replicated data and replaceable devices. The objective of this invention is providing the technique for solving the said subject.

  The data processing apparatus according to an embodiment of the present invention includes a range of sections allocated to the own apparatus among the divided sections when the management key value range, which is an m-digit continuous numerical value, is divided into a plurality of sections ( It is connected to each other data processing device (other device) to which each section other than its own range) is assigned, and is associated with a management key holding means for storing the own range and a reference key which is an m-digit numerical value. When processing data is received, if the reference key of the processing data is within the own range, the processing data is a target of the predetermined processing, and if outside the own range, the processing data is set as the reference key of the processing data. And processing data transfer means for transferring to one of other devices in normal operation to which a range including a numerical management key whose lower-order j (<m) digits match is assigned.

  The data processing program according to an embodiment of the present invention provides a range of sections assigned to its own device (within a divided range when the management key value range, which is an m-digit continuous numerical value, is divided into a plurality of sections ( A management key holding process for storing the own range in a management key holding means included in the own device in a computer connected to each other computer (other device) to which each section other than the own range is assigned, and m digits Processing data associated with a reference key that is a numerical value of the processing data, if the reference key of the processing data is within the own range, the processing data is subject to a predetermined process, and if outside the own range, Processing data transfer for transferring the processing data to any other computer in normal operation to which a range including a management key having a numerical value whose lower-order j (<m) digit matches the reference key of the processing data To perform the management.

  In the data processing method according to an embodiment of the present invention, the range of the management key value, which is an m-digit continuous numerical value, is divided into a plurality of sections. A data processing device connected to each other data processing device (other device) to which each section other than its own range) is assigned stores the own range in the management key holding means provided in the own device, and has m digits. When processing data associated with a reference key that is a numerical value is received, if the reference key of the processing data is within the own range, the processing data is subject to predetermined processing, and if the processing data is outside the own range, The processing data is transferred to one of the other devices in normal operation to which the range including the management key of the numerical value whose lower-order j (<m) digit matches the reference key of the processing data.

  The present invention makes it possible to have a suitable multiplicity of replicated data / substitutable apparatus in a system composed of a plurality of nodes according to the required reliability and the number of nodes in the system.

FIG. 1 shows the overall configuration of a data distribution processing system 10 according to the first embodiment. FIG. 2 shows an example of data storage operation of the data distribution processing system 10 of FIG. FIG. 3 shows an example of a process in which the key generation unit 24 calculates a generated key from the seed. FIG. 4 shows an example of data processing operation of the data distribution processing system 10 of FIG. FIG. 5 shows the configuration of the data processing device 20. FIG. 6 shows data stored in the management key holding unit 25. FIG. 7 is an operation flowchart of the processing data transfer unit 22. FIG. 8 is a flowchart for calculating the distance d of the processing data transfer unit 22. FIG. 9 is an operation flowchart of the key generation unit 24. FIG. 10 is an operation flowchart of the stored data transfer unit 21. FIG. 11 is an operation flowchart of the processing unit 23. FIG. 12 shows the configuration of the data processing apparatus 20 according to the second embodiment.

  FIG. 1 shows the overall configuration of a data distribution processing system 10 according to the first embodiment. The data distribution processing system 10 includes a plurality of data processing devices 20 connected to the main network 11. In FIG. 1, each of a to p surrounded by a circle is a data processing device 20. Hereinafter, the data processing device 20 represented by a in the drawing is referred to as a data processing device 20a. The same applies to b and thereafter.

  In the data distribution processing system 10, any two data processing devices 20 may not be able to communicate directly via the main network 11. Two data processing devices 20 may be able to communicate with each other via another data processing device 20.

  An administrator of the data distribution processing system 10 assigns a range of management keys to each data processing device 20. The management key is an m-bit continuous binary value. The radix of the management key may be another value. The radix is uniquely determined by the data distribution processing system 10.

  The administrator of the data distribution processing system 10 divides the range of the management key value into a plurality of continuous sections, and each section is sequentially assigned to each data processing device 20, for example, clockwise from the data processing device 20a. assign. An administrator of the data distribution processing system 10 may assign each section to each data processing device 20 and another data processing device 20 as a starting point in a counterclockwise direction.

  In FIG. 1, the data processing apparatus 20a is assigned a management key whose value ranges from sa to ea. The data processing device 20b is assigned a management key having a value ranging from sb to eb. Other data processing devices 20 are similarly assigned management key ranges.

  Here, sa is the minimum value of the management key, eq is the maximum value, and sa <ea <sb <eb. . . . <Equal to eq. One data processing device 20, for example, the data processing device 20a, may be assigned a section including the minimum value and the maximum value of the management key. In this case, the section allocated to the data processing device 20a is a combination of the section from sa to the maximum value of the management key and the section from the minimum value to eb (<sa).

  The size of the management key range to which each data processing device 20 is assigned may be the same or different. One or more management keys may be included in each range.

  At least one of the data processing devices 20, for example, the data processing device 20 j, is connected to the processing data input device 13 or the stored data input device 14 via the secondary network 12 or the like. Note that the sub network 12 and the main network 11 may be the same network.

  FIG. 2 shows an example of data storage operation of the data distribution processing system 10 of FIG. In FIG. 2, descriptions of the main network 11 and the like are omitted. An administrator of the data distribution processing system 10 assigns, for example, a management key in a range surrounded by square brackets to each data processing device 20 in the vicinity of a symbol indicating the device. For example, the data processing device 20a is assigned a management key having a value ranging from 0 to 13. The data processing device 20b is assigned a management key having a value in the range of 14 to 30. Other data processing devices 20 are similarly assigned management key ranges.

  In the data distribution processing system 10 of FIG. 2, for example, the data processing device 20j receives stored data from the stored data input device 14, and transfers the stored data to the data processing devices 20c, h, p, m. The data processing devices 20c, h, p, m receive and store the stored data.

  At this time, the data processing device 20j determines the transfer destination data processing device 20 through the process shown in FIG.

  First, the data processing device 20j obtains a reference key associated with stored data. The reference key associated with the stored data is, for example, a hash value calculated from a seed included in the stored data. The seed is data serving as a basis for calculating the hash value. The reference key may be a hash value of a seed received separately from the stored data, or may be a value of a part of the data received separately from the stored data or the stored data.

  The reference key in FIG. 3 is a hash value calculated from the seed “Tamachi Station”. The reference key is “10101010” in binary, that is, “170” in decimal.

  Next, the data processing device 20j calculates a plurality of keys (generated keys) by replacing the upper n digits of the reference key with all the values that the n digits can take.

  The generated keys in FIG. 3 are four values obtained by replacing the upper 2 bits of the hash value “10101010” with {00, 01, 10, 11}, respectively. The generated keys are decimal numbers “42”, “106”, “170”, and “234”, respectively.

  The data processing device 20j transfers the stored data to the four data processing devices 20c, h, p, m that include each generated key within the range of the assigned management key.

  FIG. 4 shows an example of data processing operation of the data distribution processing system 10 of FIG. In FIG. 4, descriptions of the main network 11 and the like are omitted. The administrator of the data distribution processing system 10 in FIG. 4 assigns, for example, management keys in the same range as in FIG. 3 to each data processing device 20.

  In the data distribution processing system 10 of FIG. 4, for example, the data processing device 20j receives the processing data from the processing data input device 13, and transfers the processing data to the data processing device 20m. The data processing device 20m receives the processing data and executes predetermined processing on the data.

  The data processing device 20j that has received the processing data calculates a generation key from the reference key associated with the processing data. This calculation is similar to the generation key calculation performed by the data processing device 20j that has received the stored data.

  For example, when the seed included in the processing data is “Tamachi Station”, as shown in FIG. 3, the generation keys are “42”, “106”, “170”, and “234” in decimal numbers, respectively. .

  The data processing device 20j includes the generated key “170” having the closest distance (see FIG. 8 etc.) with the management key range (127 to 139) assigned to the own device within the assigned management key range. The processing data is transferred to the data processing device 20m to be included.

  The processing data is transferred hop-by-hop from the data processing device 20j to the data processing device 20m. The processing data may pass through another data processing device 20 such as the data processing device 20k or the data processing device 20l before reaching the data processing device 20m. Which data processing device 20 is routed depends on the routing protocol applied to the data distribution processing system 10.

  However, in the data distribution processing system 10 according to the present embodiment, the number of pops between the data processing devices 20 with the assigned management key distance is less than the number of pops between the data processing devices 20 with the assigned management key distance far. Use a routing protocol such that An example of such a protocol is described in Non-Patent Document 2.

  Predetermined processing executed by the data processing device 20m is determined corresponding to the data distribution processing system 10.

  The reason why the processing data received by the data processing device 20j is transferred to the data processing device 20m is, for example, that the predetermined processing requests the storage data described in FIG. Alternatively, for example, the predetermined processing requires a device such as a high-performance processor or an image processing engine provided only in the data processing devices 20c, h, p, and m.

  If the data processing device 20m does not operate normally, the processing data is transferred to the data processing device 20p, for example. The reason is that the data processing device 20p has a copy of the same stored data as the data processing device 20m as shown in FIG. 2, or a device that can replace the device provided in the data processing device 20m.

  The data distribution processing system 10 of this embodiment operates as a publish / subscribe (pub / sub) type system, for example. In this system, an application program executed by the stored data input device 14 or the like inputs a subscription as stored data in advance to any data processing device 20, for example, the data processing device 20j. The subscription is stored in any data processing device 20, for example, the data processing device 20m.

  The subscription is composed of an event condition and an address of an external system that is notified when an event that satisfies the condition is input.

  When a certain data processing device 20 inputs an event as processing data from the processing data input device 13, the data processing device 20 transfers the event to the data processing device 20 that stores the subscription. For example, the data processing device 20j inputs an event as processing data from the processing data input device 13, and transfers the event to the data processing device 20m that stores the subscription.

  The data processing device 20 that stores the subscription collates the transferred event with the event condition of the registered subscription, and if it matches, notifies the external system designated by the subscription of the event.

  In this case, the collation of the input event with the registered event condition and the notification to the external system correspond to the predetermined processing described above.

  An event is processing data composed of one or more data items. Each data item is composed of an attribute name and an attribute value. For example, when the processing data input device 13 generates an event based on the temperature of 20 ° C. and the humidity of 60% observed by the sensor, the event is a data item whose attribute name is “temperature” and whose attribute value is 20 ° C. And an attribute name “humidity” and an attribute value of 60%.

  The data processing apparatus 20 of this system calculates a generation key using the subscription and event attribute values as seeds.

  The data distribution processing system 10 of this embodiment can also function as a storage that stores object data in a distributed manner. In this case, each object data is stored data that each data processing device 20 shares and stores.

  A processing data input device 13 that requests object data inputs a data condition (query) to the data distribution processing system 10 as processing data. The data processing device 20, for example, the data processing device 20j, to which the query is input, specifies the data processing device 20, for example, the data processing device 20m, that stores the object data in the same manner as described above.

  The data processing device 20j receives the object data as a query execution result from the data processing device 20m, and responds to the processing data input device 13. In this case, executing the query to obtain the requested object data is a predetermined process executed by the data processing device 20.

  FIG. 5 shows the configuration of the data processing apparatus 20 of the present embodiment. The data processing device 20 includes a processing data transfer unit 22 and a management key holding unit 25. The data processing device 20 may include a stored data transfer unit 21, a processing unit 23, a key generation unit 24, a communication unit 27, and a stored data holding unit 26.

  The stored data transfer unit 21 receives the stored data from the stored data input device 14 and transfers it to another data processing device 20 (see FIG. 2). The processing data transfer unit 22 receives the processing data from the processing data input device 13 and transfers it to another data processing device 20 (see FIG. 4). The key generation unit 24 calculates a generation key from the seed (see FIG. 3).

  The processing unit 23 has functions such as inputting processing data and executing predetermined processing. The communication unit 27 communicates with other data processing devices 20. The communication unit 27 may execute routing protocol processing.

  The stored data transfer unit 21 and the processing data transfer unit 22 may include the function of the key generation unit 24. The stored data transfer unit 21, the processing data transfer unit 22, and the processing unit 23 may include the function of the communication unit 27. Furthermore, the processing unit 23 and the stored data holding unit 26 may exist in another device connected to the data processing device 20, for example, a database server.

  The stored data transfer unit 21, the processing data transfer unit 22, and the processing unit 23 are realized by dedicated hardware. All or part of the stored data transfer unit 21, the processing data transfer unit 22, and the processing unit 23 may be realized by executing a data processing program 28 by a processor (not shown) of the data processing device 20.

  The management key holding unit 25 and the stored data holding unit 26 are storage devices such as a disk device and a semiconductor memory. The stored data holding unit 26 stores stored data such as the above-described subscription and object data. The specific data stored data depends on the data distribution processing system 10.

  FIG. 6 shows data stored in the management key holding unit 25. An administrator of the data distribution processing system 10 sets values in the management key holding unit 25 for the own range 30, the key length 31, and the number of duplicate digits 32.

  The own range 30 holds the minimum value and the maximum value of the management key range assigned to the own device. For example, in the data processing device 20j shown in FIGS. 2 and 3, kmin is 127 and kmax is 139. The stored data held by the stored data holding unit 26 is such that the generated key k calculated from the seed is within the range of the own range 30.

  The key length 31 holds the number m (positive integer) of the management key, the reference key, and the generated key. In the example of FIG. 3, m is 8 bits. The key generation unit 24 generates a key having a length specified by the key length 31.

  The number of duplicate digits 32 holds the upper limit of the number of data processing devices 20 (duplication multiplicity) for storing duplicates of stored data in the form of the number n of upper digits of the reference key and the generated key. The upper limit of the multiplicity of replication is the nth power of the base of the reference key and the generated key. In the example of FIG. 3, n is 2 bits, and the upper limit of the replication multiplicity is 2 squared to 4.

  In the present invention, there are cases where the m-digit reference key and the generated key are divided into upper n digits and lower j (= mn) digits.

  Note that the replication multiplicity of stored data in the data distribution processing system 10 may be smaller than the upper limit. The stored data transfer unit 21 calculates a generation key having an upper limit number of distributed values, and tries to create a copy in the data processing apparatus 20 including each generation key in its own range 30. However, a plurality of calculated generated keys may be included in the own range 30 of one data processing apparatus 20, and in this case, the replication multiplicity is smaller than the upper limit.

  FIG. 7 is an operation flowchart of the processing data transfer unit 22. When the process data transfer unit 22 receives the process data from the process data input device 13 (S1), the process data transfer unit 22 inputs a seed of the process data to the key generation unit 24 and acquires one or more generation keys from the key generation unit 24. (S2).

  Next, the processing data transfer unit 22 obtains a minimum distance dmin and a transfer target key that is a generated key that gives the dmin out of the range of the management key assigned to the own device and each generated key. .

  First, the processing data transfer unit 22 initializes dmin as MAX (S3). MAX is the maximum value that the generated key can take. When the key length 31 is m, the maximum value of the generated key is (radix m-1).

  Next, the process data transfer unit 22 repeatedly executes the processes of steps S4 to S6 for each generated key k.

  The process data transfer unit 22 executes the process shown in FIG. 8 to be described later to obtain the distance d between the own range 30 and k (S4). If dmin> d (Y in S5), the process data transfer unit 22 newly sets d as dmin and k as a transfer target key (S6).

  If dmin = 0 is not satisfied after completion of the above-described processing for each generated key k (N in S7), the processing data transfer unit 22 transfers the processing data to the data processing device 20 to which the transfer target key is assigned (S8). .

  If the transfer is completed normally (Y in S9), the process data transfer unit 22 ends the process. If the transfer is not completed normally (N in S9), the process data transfer unit 22 deletes the generated key that is the current transfer target key (SA) and repeats the process from S3.

  If the transfer is not completed normally (N in S9), the processing data transfer unit 22 determines that a failure has occurred in the data processing device 20 that is the transfer destination indicated by the current transfer target key and performs the above processing. . Note that the generated key deleted in step SA is not a processing target in steps S4 to S6.

  If dmin = 0 (Y in S7), the processing data transfer unit 22 transmits the processing data to the processing unit 23 of its own device (SB) and ends the processing.

  FIG. 8 is a flowchart for calculating the distance d of the processing data transfer unit 22. This figure is a flowchart for calculating a distance (clockwise distance) on the assumption that each data processing device 20 in FIG. 2 or 3 transfers the processing data in the clockwise direction. For example, when the protocol described in Non-Patent Document 2 is used, each data processing device 20 transfers the processing data in the clockwise direction.

  When kmin <k <kmax holds for the generated key k (Y in S11), the processing data transfer unit 22 sets the distance d to 0 (S12). kmin and kmax are values held by the own range 30.

  When kmin <k <kmax is not satisfied (N in S11) and k> kmax is satisfied (Y in S14), the processing data transfer unit 22 sets the distance d to k−kmax (S13). If kmin <k <kmax is not satisfied (N in S11) and k> kmax is not satisfied (N in S14), the processing data transfer unit 22 sets the distance d to MAX−kmax + k + 1 (S15).

  When the protocol described in Non-Patent Document 2 is used, each data processing device 20 is separated by a power of 2 (i−1) from the maximum value of the management key assigned to the data processing device 20 at a clockwise distance. The address of the data processing device 20 to which the management key is assigned is held. As a result, the hop count of the processing data in the data distribution processing system 10 is on the order of log (number of data processing devices / 2 to the nth power).

  When each data processing device 20 receives the processing data, each data processing device 20 uses the data processing device 20 assigned the management key closest to the reference key of the processing data as the transfer destination. The transfer data reaches the data processing device 20 that manages the key on a hop-by-hop basis.

  The distance between the keys is not limited to the clockwise distance. For example, the processing data transfer unit 22 may use a counterclockwise distance as the distance between the keys. In this case, the distance d is k> kmin in S14 of FIG. 8, MAX−k + kmin + 1 in S13, and Kmin−k in S15. Further, the processing data transfer unit 22 may set the distance between keys to a smaller value between the clockwise distance and the counterclockwise distance.

  The administrator of the data distribution processing system 10 or the like, the number of pops between the data processing devices 20 with a short distance between keys is equal to or less than the number of pops between the data processing devices 20 with a long distance according to a routing algorithm executed by the system. Define the distance to be The processing data transfer unit 22 calculates the distance according to the definition.

  Further, the processing data transfer unit 22 may consider the priority in addition to the above distance when selecting the transfer destination. This is effective when, for example, a plurality of replaceable devices included in the data distribution processing system 10 are not equivalent. In this case, for example, the processing data transfer unit 22 weights and adds the above-described distance and priority, and sets the data processing device 20 that gives the minimum value of the addition value as the transfer destination.

  FIG. 9 is an operation flowchart of the key generation unit 24. When receiving the seed (S21), the key generation unit 24 first inputs the seed into a hash function to obtain a hash value (S22). Next, the key generation unit 24 acquires the number of bits n from the number of duplicate digits 32 (S23), and replaces the upper n bits of the hash value with all possible combinations of the n bits to generate keys ( S24). As described above, FIG. 3 shows a generation key calculation example of the key generation unit 24.

  FIG. 10 is an operation flowchart of the stored data transfer unit 21. When the storage data transfer unit 21 receives the storage data and the request type from the storage data input device 14 (S31), the storage data transfer unit 21 inputs the seed of the storage data to the key generation unit 24 and acquires one or more generation keys (S32). .

  Next, the stored data transfer unit 21 repeatedly executes the processes of steps S33 to S35 for each generated key k.

  When kmin <k <kmax does not hold for the generated key k (N in S33), the stored data transfer unit 21 transfers the stored data and the request type to the data processing device 20 to which the generated key k is assigned (S34). . kmin and kmax are values held by the own range 30.

  When kmin <k <kmax is satisfied (Y in S33), the stored data transfer unit 21 transmits the stored data and the request type to the processing unit 23 of the own apparatus (S35).

  The request type is an identifier for identifying whether stored data is registered or deleted, and has a value of “registration” or “deletion”.

  FIG. 11 is an operation flowchart of the processing unit 23. The processing unit 23 receives stored data from the stored data transfer unit 21 of another data processing device 20 or its own device. The processing unit 23 receives processing data from the processing data transfer unit 22 of another data processing device 20 or its own device.

  If the received data is processed data (Y in S41), the processing unit 23 executes a predetermined process on the processed data (S44) and ends the operation. If the received data is stored data (N in S41), the processing unit 23 executes processing according to the request type and ends the operation.

  When the request type is “registration” (Y in S42), the processing unit 23 stores the received storage data in the storage data holding unit 26 (S43). When the request type is “delete” (N in S42), the processing unit 23 deletes data that matches the received stored data from the stored data holding unit 26 (S45).

  The data distribution processing system 10 according to the present embodiment is a system including a plurality of data processing devices 20. In the data processing apparatus 20 of the present embodiment, the data distribution processing system 10 has appropriate multiplicity of replicated data / substitutable apparatuses according to the required reliability and the number of data processing apparatuses 20 in the system. Is possible.

  The reason is that the processing data transfer unit 22 transfers the processing data to any other device in normal operation to which a range including a management key whose lower-order j digit matches the reference key of the processing data is assigned. It is.

  An administrator or the like of the data distribution processing system 10 can set an appropriate own range 30 and duplication digit number 32 after arranging stored data and replaceable devices according to a desired multiplicity.

  In the data processing apparatus 20 according to the present embodiment, the data distribution processing system 10 further creates a copy data with appropriate multiplicity according to the required reliability and the number of data processing apparatuses 20 in the system. Make it possible.

  The reason is that the stored data transfer unit 21 transfers the stored data to a plurality of other devices in normal operation to which a range including a management key whose lower-order j digit matches the reference key of the processing data is assigned. is there.

  An administrator of the data distribution processing system 10 can set an appropriate number of duplication digits 32 according to the desired multiplicity.

  FIG. 12 shows the configuration of the data processing apparatus 20 according to the second embodiment of the present invention. The data processing device 20 has a range other than the range (own range 30) allocated to the own device among the divided sections when the management key value range, which is an m-digit continuous numerical value, is divided into a plurality of sections. Each section is connected to each other data processing apparatus 20 (other apparatus) to which each section is assigned.

  The data processing device 20 includes a management key holding unit 25 that stores its own range 30 and a processing data transfer unit 22.

  The processing data transfer unit 22 receives the processing data associated with the reference key that is an m-digit numerical value, and if the processing data reference key is within its own range 30, the processing data is subject to predetermined processing. Further, if the reference key of the processing data is outside the range 30, the processing data transfer unit 22 has a range that includes the management key with a numerical value that matches the processing data reference key and the lower j (<m) digits. Transfer to one of the other assigned devices in normal operation.

  In the data processing apparatus 20 of the present embodiment, the data distribution processing system 10 has appropriate multiplicity of replicated data / substitutable apparatuses according to the required reliability and the number of data processing apparatuses 20 in the system. Is possible.

  The reason is that the processing data transfer unit 22 transfers the processing data to any other device in normal operation to which a range including a management key whose lower-order j digit matches the reference key of the processing data is assigned. It is.

DESCRIPTION OF SYMBOLS 10 Data distributed processing system 11 Main network 12 Sub network 13 Process data input device 14 Stored data input device 20 Data processor 21 Stored data transfer part 22 Process data transfer part 23 Process part 24 Key generation part 25 Management key holding part 26 Stored data Holding unit 27 Communication unit 28 Data processing program 30 Own range 31 Key length 32 Number of replicated digits

Claims (8)

Each assigned to each section other than the section range (own range) assigned to the own device among the divided sections when the range of the management key value, which is an m-digit continuous value, is divided into a plurality of sections. Connected to other data processing devices (other devices)
Management key holding means for storing the own range;
Receive processing data associated with the reference key, which is an m-digit number,
If the reference key of the processing data is outside the range, the processing data is
(A) The smallest management among other devices in normal operation to which a range including a management key having a numerical value larger than the own range, in which the lower-order j (<m) digit matches the reference key of the processing data. Other device (high-order device) assigned a key range, or
(B) The maximum management among other devices in normal operation to which a range including a management key with a numerical value smaller than the own range, in which the reference key of the processing data matches the lower j (<m) digits, is assigned. Other device (subordinate device) assigned key range,
Data processing apparatus comprising a processing data transfer means for transferring the one of. It said processing data transfer means, the processing data, among the upper device and said lower device, and transfers the one constant order was, or,
The distance between the own range and the range of the management key assigned to the higher-level device is compared with the distance between the own range and the range of the management key assigned to the lower-level device, and the distance is transferred to the smaller one. Data processing equipment. Stored data holding means;
Processing means for executing the given processing using data stored in the storage data holding means,
If the stored data associated with the reference key is received, and the key whose lower j (<m) digit matches the reference key of the stored data is within the own range, the processed data is stored in the stored data holding means. Store and
2. A stored data transfer means for transferring the stored data to a plurality of other devices to which a range including a management key whose lower n digits match the reference key of the stored data is assigned if it is outside the range. The data processing device according to any one of 1 to 2 . A data processing device according to any one of claims 1 to 3 ,
Each other device assigned each section other than the own range;
A data distribution processing system including a processing data input device connected to the data processing device and transmitting processing data to the data processing device. Each assigned to each section other than the section range (own range) assigned to the own device among the divided sections when the range of the management key value, which is an m-digit continuous value, is divided into a plurality of sections. To computers connected to other computers (other devices)
A management key holding process for storing the own range in a management key holding means provided in the own device;
Receive processing data associated with the reference key, which is an m-digit number,
If the reference key of the processed data is outside the own range, the pre-Symbol processing data,
(A) The smallest management among other devices in normal operation to which a range including a management key having a numerical value larger than the own range, in which the lower-order j (<m) digit matches the reference key of the processing data. Other device (high-order device) assigned a key range, or
(B) The maximum management among other devices in normal operation to which a range including a management key with a numerical value smaller than the own range, in which the reference key of the processing data matches the lower j (<m) digits, is assigned. Other device (subordinate device) assigned key range,
It said processing Lud over data processing program to execute the data transfer process of transferring the one of. In the computer,
Said process data, among the upper device and said lower device, and transfers one was constant because, or,
The processing data transfer in which the distance between the own range and the range of the management key assigned to the higher-level device and the distance between the own range and the range of the management key assigned to the lower-level device are compared and transferred to one having a smaller distance The data processing program according to claim 5 , wherein the processing is executed. In the computer,
And processing execution processing for executing the given process using the stored data in the storage data holding means included in the computer,
If the stored data associated with the reference key is received, and the key whose lower j (<m) digit matches the reference key of the stored data is within the own range, the processed data is stored in the stored data holding means. Store and
A storage data transfer process is executed for transferring the stored data to a plurality of other devices to which a range including a management key whose lower-order n digits match the reference key of the stored data is outside the own range. A data processing program according to any one of 5 to 6 . Each assigned to each section other than the section range (own range) assigned to the own device among the divided sections when the range of the management key value, which is an m-digit continuous value, is divided into a plurality of sections. A data processing device connected to another data processing device (other device)
The own range is stored in the management key holding means provided in the own device,
Receive processing data associated with the reference key, which is an m-digit number,
If the reference key of the processed data is outside the own range, the process data
(A) The smallest management among other devices in normal operation to which a range including a management key having a numerical value larger than the own range, in which the lower-order j (<m) digit matches the reference key of the processing data. Other device (high-order device) assigned a key range, or
(B) The maximum management among other devices in normal operation to which a range including a management key with a numerical value smaller than the own range, in which the reference key of the processing data matches the lower j (<m) digits, is assigned. Other device (subordinate device) assigned key range,
Data processing method to transfer to one of the.

Images