JP5272927B2 - Node device, program, and distributed storage method - Google Patents

Description

  The present invention relates to a technical field of a peer-to-peer (P2P) type communication system including a plurality of node devices that can communicate with each other via a network.

  This type of peer-to-peer communication system (P2P network) distributes and stores (arranges) duplicate data files of a plurality of data files in a plurality of node devices. Each data file is a data file of various contents such as movies and music. A distributed storage system in which the above-described data file can be used between node devices is known. This improves fault tolerance and access dispersibility. The location of the replicas thus stored in a distributed manner can be efficiently searched using a distributed hash table (hereinafter referred to as DHT (Distributed Hash Table)) as disclosed in Patent Document 1.

JP 2006-197400 A

  By the way, such a distributed storage system generally includes a management server. The management server manages information indicating a plurality of node devices in which a plurality of data files are stored. Specifically, the management server again distributes and saves the data file so that the data file is not lost from the system when the node device fails. The management server redistributes the data file when a new node device joins the system.

  In this case, each node device periodically notifies the management server of the data file stored in each node device. When the management server grasps the data file of each node device, storage control of the data file can be performed.

  However, such a method places an excessive processing burden on the management server. As the number of data files provided in the distributed storage system increases and the number of node devices participating in the distributed storage system increases, the processing load on the management server increases.

  The present invention has been made in view of the above problems and the like, and provides a node device, a processing program, and a distributed storage method in which a load applied during the data file storage control is distributed to a plurality of node devices. Is an issue.

In order to solve the above-mentioned problem, according to the first aspect of the present invention, a copy number corresponding to a predetermined number of copies is assigned to a data file that can be transmitted and received between a plurality of node devices that can communicate with each other via a network. Replicating as a replica, the data file corresponding to the replicated replica can be identified, and for the plurality of identification information to be assigned to the data file, the plurality of identification information is distributed and stored in the plurality of node devices First determination means for determining a part of the plurality of identification information at a predetermined replication number in the predetermined number of replicas as a storage charge range, and a replication number determined by the first determination means; The storage responsibility range for each different replication number is different for each of the replication numbers for at least the storage responsibility range determined by the first determination means. A second determining means for determining by the storage control means for storing the replica storage means of said first storage charge range determined with decision means by said second determining means, a total number of said plurality of identification information Total number acquisition means to acquire, apparatus number acquisition means to acquire the number of node devices indicating the number of the plurality of node devices, and the leaving node from the leaving node device that leaves the distributed storage system configured by the plurality of node devices In the case of including a leave information receiving unit that receives device leave information, or a leave message receiving unit that receives a leave message of the leave node device from another node device that has received the leave information, and the leave information receiving unit. When the leaving information is received, or when the leaving message receiving means is provided, the leaving message is received. A node device number updating means for updating the number of node devices based on the leaving information or the leaving message, wherein the first determining means updates the number of node devices updated by the node device number updating means. And the total number and the total number of node devices in the predetermined replication number are determined again to be equal, and the second determining unit is updated by the node device number updating unit. Based on the number of node devices and the total number, the storage responsibility range of the plurality of node devices for each of the different replication numbers is re-determined so that the storage control unit has the first determination unit and the A node device characterized in that a storage unit stores a replica of a storage charge range re-determined by a second determination unit .

In order to solve the above-mentioned problem, the invention according to claim 2 assigns a copy number corresponding to a predetermined number of copies to a data file that can be transmitted and received between a plurality of node devices that can communicate with each other via a network. Replicating as a replica, the data file corresponding to the replicated replica can be identified, and for the plurality of identification information to be assigned to the data file, the plurality of identification information is distributed and stored in the plurality of node devices First determination means for determining a part of the plurality of identification information at a predetermined replication number in the predetermined number of replicas as a storage charge range, and a replication number determined by the first determination means; The storage responsibility range for each different replication number is different for each of the replication numbers for at least the storage responsibility range determined by the first determination means. Second determining means for determining, storage control means for storing in the storage means a replica of the storage charge range determined by the first determining means and the second determining means, and a total number of the plurality of identification information From the total number acquisition means to acquire, the device number acquisition means for acquiring the number of node devices indicating the number of the plurality of node devices, and a new node device newly participating in the distributed storage system constituted by the plurality of node devices, A participation message receiving means for receiving the participation message; and a node device number updating means for updating the number of node devices based on the participation message when the participation message is received by the participation message receiving means, The first determining unit is configured to perform the predetermined complex based on the number of node devices updated by the node device number updating unit and the total number. Re-determined so that the storage charge range for each of the plurality of node devices in the number is equal, the second determining means, based on the number of node devices updated by the node device number updating means and the total number, Re-determining the storage responsibility ranges of the plurality of node devices for each different replication number to be equal, and the storage control means re-determining the storage responsibility ranges re-determined by the first determination means and the second determination means The replica is stored in a storage means .

In order to solve the above-mentioned problem, the invention according to claim 3 assigns a copy number corresponding to a predetermined number of copies to a data file that can be transmitted and received between a plurality of node devices that can communicate with each other via a network. Replicating as a replica, the data file corresponding to the replicated replica can be identified, and for the plurality of identification information to be assigned to the data file, the plurality of identification information is distributed and stored in the plurality of node devices A first determination step of determining a part of the plurality of identification information at a predetermined replication number in the predetermined number of replicas as a storage charge range, and a replication number determined by the first determination step; Is a storage responsibility range for each different replication number, at least for each of the storage responsibility ranges determined by the first determination step. A second determination step that is determined differently for each storage, a storage control step for storing in the storage means a replica of the storage charge range determined by the first determination step and the second determination step, and the plurality of identification information A total number obtaining step for obtaining the total number of nodes, a device number obtaining step for obtaining the number of node devices indicating the number of the plurality of node devices, and a leaving node device leaving the distributed storage system configured by the plurality of node devices. A leave information receiving step for receiving leave information of the leave node device, or a leave message receiving step for receiving a leave message of the leave node device from another node device that has received the leave information, and the leave information receiving step Or when the leaving information is received by the leaving message receiving step. When the leaving message is received, a program for causing a computer to execute a node device number updating step of updating the number of node devices based on the leaving information or the leaving message, wherein the first determination Step is re-determined based on the number of node devices updated by the node device number update step and the total number, so that the storage responsibility range for each of the plurality of node devices in the predetermined replication number is equal, Based on the number of node devices updated by the node device number updating step and the total number, the second determining step is repeated so that storage ranges of the plurality of node devices are equalized for each of the different replication numbers. And the storage control step is redetermined by the first determination step and the second determination step. A program for causing a computer to store a replica of a storage charge range in a storage means .

In order to solve the above-described problem, the invention according to claim 4 assigns a copy number corresponding to a predetermined number of copies to a data file that can be transmitted and received between a plurality of node devices that can communicate with each other via a network. Replicating as a replica, the data file corresponding to the replicated replica can be identified, and for the plurality of identification information to be assigned to the data file, the plurality of identification information is distributed and stored in the plurality of node devices A first determination step of determining a part of the plurality of identification information at a predetermined replication number in the predetermined number of replicas as a storage charge range, and a replication number determined by the first determination step; Is a storage responsibility range for each different replication number, at least for each of the storage responsibility ranges determined by the first determination step. A second determination step that is determined differently for each storage, a storage control step for storing in the storage means a replica of the storage charge range determined by the first determination step and the second determination step, and the plurality of identification information A total number acquisition step of acquiring the total number of nodes, a device number acquisition step of acquiring the number of node devices indicating the number of the plurality of node devices, and a new node newly participating in the distributed storage system configured by the plurality of node devices A participation message receiving step for receiving the participation message from a device; and a node device number updating step for updating the number of node devices based on the participation message when the participation message is received by the participation message receiving step; Is a program for causing a computer to execute the first determination step. Is re-determined based on the number of node devices updated in the node device number update step and the total number so that the storage responsibility ranges for the plurality of node devices at the predetermined replication number are equalized, Based on the number of node devices updated by the node device number updating step and the total number, the second determining step is repeated so that storage ranges of the plurality of node devices are equalized for each of the different replication numbers. The storage control step is a program for causing a computer to store a replica of a storage charge range re-determined by the first determination step and the second determination step in a storage unit. .

In order to solve the above-mentioned problem, the invention according to claim 5 assigns a copy number corresponding to a predetermined number of copies to a data file that can be transmitted and received between a plurality of node devices that can communicate with each other via a network. The node device is capable of identifying a data file corresponding to the replicated replica as a replica, and the plurality of identification information to be assigned to the data file. A first determination step of determining, as a storage charge range, a part of the plurality of identification information at a predetermined replication number of the predetermined number of replicas, and the node device includes: A storage responsibility range for each copy number different from the copy number determined in one determination step is determined by at least the first determination means. A second determination step for determining differently for each of the replication numbers for the determined storage assigned range, and a replica of the storage assigned range determined by the node device by the first determining step and the second determining step. A storage control step for storing in the storage means; a total number acquisition step for acquiring a total number of the plurality of identification information; a device number acquisition step for acquiring a node device number indicating the number of the plurality of node devices; and the plurality of nodes A leaving information receiving step for receiving leaving information of the leaving node device from a leaving node device that leaves the distributed storage system configured by the device, or leaving the leaving node device from another node device that has received the leaving information The leaving information is received by the leaving message receiving step for receiving a message and the leaving information receiving step. Or a node device number updating step of updating the number of node devices based on the leave information or the leave message when the leave message is received by the leave message receiving step. 1 determination step is re-determined based on the number of node devices updated by the node device number update step and the total number so that the storage responsibility ranges for the plurality of node devices at the predetermined replication number are equalized. In the second determination step, the storage responsibility ranges of the plurality of node devices for the different replication numbers are equalized based on the number of node devices updated by the node device number updating step and the total number. The storage control step is redetermined by the first determination step and the second determination step. The distributed storage method is characterized in that the storage means replica is stored in the storage means .

In order to solve the above-mentioned problem, the invention according to claim 6 assigns a copy number corresponding to a predetermined number of copies to a data file that can be transmitted and received between a plurality of node devices that can communicate with each other via a network. The node device is capable of identifying a data file corresponding to the replicated replica as a replica, and the plurality of identification information to be assigned to the data file. A first determination step of determining, as a storage charge range, a part of the plurality of identification information at a predetermined replication number of the predetermined number of replicas, and the node device includes: A storage responsibility range for each copy number different from the copy number determined in one determination step is determined by at least the first determination means. A second determination step for determining differently for each of the replication numbers for the determined storage assigned range, and a replica of the storage assigned range determined by the node device by the first determining step and the second determining step. A storage control step for storing in the storage means; a total number acquisition step for acquiring a total number of the plurality of identification information; a device number acquisition step for acquiring a node device number indicating the number of the plurality of node devices; and the plurality of nodes A participation message receiving step for receiving the participation message from a new node device newly participating in the distributed storage system constituted by the devices, and when the participation message is received by the participation message receiving step, based on the participation message And updating the number of node devices, and updating the number of node devices. Step is re-determined based on the number of node devices updated by the node device number update step and the total number, so that the storage responsibility range for each of the plurality of node devices in the predetermined replication number is equal, Based on the number of node devices updated by the node device number updating step and the total number, the second determining step is repeated so that storage ranges of the plurality of node devices are equalized for each of the different replication numbers. The distributed storage method is characterized in that the storage control step stores the storage unit replica of the storage charge range re-determined by the first determination step and the second determination step .

  According to the present invention, each node device can store a plurality of replicas redundantly, and each node device itself can autonomously perform storage control of a replica in charge of storage without using a management server. Can do.

It is a figure which shows an example of the connection aspect of each node apparatus in the distributed storage system S which concerns on this embodiment. It is explanatory drawing of node ID and file ID in ID space. It is explanatory drawing of the storage charge range of each node Nn in each hierarchy in case a hierarchy number is "2". (A) is explanatory drawing of the memory charge range of each node Nn in a 1st hierarchy. (B) is an explanatory diagram of a storage charge range of each node Nn in the second hierarchy. (A) is explanatory drawing of the memory | storage charge range in a 1st hierarchy (FIG. 3 (A)). (B) is an explanatory diagram of a storage charge range in the second hierarchy (FIG. 3 (B)). It is explanatory drawing which shows the mode that the new node N5 connects to the distributed storage system S. It is explanatory drawing of the storage charge range of each node Nn in each hierarchy in case a hierarchy number is "3". (A) is explanatory drawing of the memory charge range of each node Nn in a 1st hierarchy. (B) is an explanatory diagram of a storage charge range of each node Nn in the second hierarchy. (C) is an explanatory diagram of a storage charge range of each node Nn in the third hierarchy. It is a block diagram which shows the example of a schematic structure of the node Nn. It is a processing flow when a new node Nn connects to the distributed storage system S, (A) is a process of the control unit 11 in the new node Nn, and (B) is already connected to the distributed storage system S. This is a process of the control unit 11 in the node Nn. The processing flow of each node Nn when a leaving node is generated from the distributed storage system S, (A) is the processing of the control unit 11 in the node Nn that has received leaving information indicating that it is leaving from the leaving node, (B) is a process of the control unit 11 in the node Nn that has received the leave message. It is explanatory drawing of the determination method of the storage charge range of the 2nd hierarchy in a modification, (A) is explanatory drawing of the storage charge range of each node Nn in the 1st hierarchy of the determination methods 1, 2, and 3. FIG. (B) is an explanatory diagram of the storage responsibility range of each node Nn in the second hierarchy in the case of the determination method 1. (C) is an explanatory diagram of the storage charge range of each node Nn in the second hierarchy in the case of the determination method 2. FIG. 6D is an explanatory diagram of a storage charge range of each node Nn in the second hierarchy in the case of the determination method 3.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

1. Configuration and Operation Overview of Distributed Storage System First, the configuration and operation overview of the distributed storage system according to the present embodiment will be described with reference to FIG.

  FIG. 1 is a diagram illustrating an example of a connection mode of each node device in the distributed storage system according to the present embodiment.

  As shown in the lower frame 101 of FIG. 1, the distributed storage system S is configured by connecting a plurality of node devices Nn via the Internet. Further, as shown in the lower frame 101 of FIG. 1, IX (Internet eXchange) 3, ISP (Internet Service Provider) 4a, 4b, DSL (Digital Subscriber Line) line operators (devices) 5a, 5b, FTTH ( A network 8 is constructed by a fiber operator (device) 6 and a communication line 7. The communication line 7 is a telephone line or an optical cable. The network 8 is a real world network such as the Internet. A router for transferring data (packets) is appropriately inserted in the network 8 in the example of FIG.

  A plurality of node devices Nn (n = 1, 2, 3,...) Are connected to such a network 8. Hereinafter, the node device is referred to as a “node”. Each node Nn is assigned a unique manufacturing number and an IP (Internet Protocol) address. The distributed storage system S according to the present embodiment includes a peer-to-peer network system formed by connecting any of a plurality of nodes Nn as shown in the upper frame 100 of FIG. It has become.

  The network 9 shown in the upper frame 100 of FIG. 1 is an overlay network 9 that configures a virtual link formed using the existing network 8. The overlay network 9 is a logical network and is realized by a specific algorithm, for example, an algorithm using DHT.

  A node ID is assigned to each node Nn connected to the distributed storage system S, in other words, connected to the overlay network 9. The node ID is unique identification information obtained by a node ID determination method described later.

  Each node Nn holds a routing table using DHT. This routing table defines various message transfer destinations on the distributed storage system S. Specifically, the routing table registers a plurality of pieces of node information including node IDs and address information of nodes Nn that are moderately separated in the ID space. The address information includes an IP address and a port number. By communicating with each other, each node Nn can know each other's node information via another node Nn, or via a management server or the like.

  By the way, in the distributed storage system S, replicas, which are replicated data of various data files having different contents, are distributed (stored) in a plurality of nodes Nn in a predetermined file format. The replica can be used between the nodes Nn. For example, the node N5 stores a replica of a movie data file whose title is XXX as an example of a replica. On the other hand, a replica of the data file of the movie whose title is YYY is stored in the node N3.

  In addition, information such as a file ID and a data file name is added to each of these replicas. This file ID is unique identification information for each data file.

  File IDs are generated so as to be probabilistically equal in the ID space and assigned to each replica. A predetermined number of replicas of each data file are generated. In the following description, the number of copies may be referred to as a duplication number. The same file ID is assigned to replicas of the same data file. The file ID is generated, for example, by hashing “data file name + arbitrary numerical value” with a hash function such as SHA-1. Further, the first several bytes of data constituting the data file may be generated by hashing. Alternatively, the system administrator may randomly assign ID values that are unique ID values for each data file and are stochastically equal in the ID space. The file ID has the same bit length as the node ID so that the file ID is defined in the same ID space as the node ID. The bit length is, for example, 160 bits. For example, both the file ID and the node ID are expressed by hexadecimal numbers of 4 digits. Such a replica of the data file is stored in the node Nn in charge of storage determined based on a predetermined rule as a person in charge of storage based on each file ID.

  In the present embodiment, the same data file replica is stored in a plurality of nodes Nn in an overlapping manner.

1-1. Method for Determining Node ID and Storage Range The method for determining the node ID and storage range of each node Nn will be described.

  FIG. 2 is an explanatory diagram of a node ID and a file ID in the ID space. In the figure, a white circle (◯) indicates a node ID, and a black circle (●) indicates a file ID.

  FIG. 2 shows an example in which four nodes Nn, that is, a node N1, a node N2, a node N3, and a node N4 are connected to the distributed storage system S.

  In the present embodiment, each node ID and file ID is arranged in an ID space constituting a 4-digit hexadecimal number (0000 to ffff). The total number of ID spaces is 10,000 (１０ffff + 1).

  In the example shown in FIG. 2, the replica storage range is determined by equally dividing the total number of ID spaces 10000 by the nodes N1 to N4. Each replica is stored in a node Nn in charge of storage determined based on a predetermined rule. In the present embodiment, it is assumed that the node ID is stored in a node Nn having a node ID equal to or less than the file ID and having the smallest difference between the file ID and the node ID.

In the example shown in FIG.
The node ID of the node N1 is “0000”, and the node N1 stores a replica having a file ID in the ID (0000-3fff) area in the ID space. For example, a replica such as a file ID “2fff” is stored.

  The node ID of the node N2 is “4000”, and the node N2 stores a replica having a file ID in the ID (4000 to 7fff) area in the ID space. For example, a replica such as a file ID “6500” is stored.

  The node ID of the node N3 is “8000”, and the node N3 stores a replica having a file ID in the ID (8000 to bfff) area in the ID space. For example, a replica such as a file ID “9000” is stored.

  The node ID of the node N4 is “c000”, and the node N4 stores a replica having a file ID in the ID (c000 to ffff) area in the ID space. For example, a replica such as a file ID “d000” is stored.

  In this way, each of the nodes N1 to N4 stores a replica of each storage charge range.

  Furthermore, in this embodiment, the storage charge ranges of the nodes N1 to N4 are defined in a plurality of ID space layers so that the replicas of the same data file are stored in the plurality of nodes Nn. The number of hierarchies “k” is conditional on being a natural number of 2 or more. In addition, the storage charge range of each node Nn is required to be less than 1 / hierarchy. The first hierarchy, the second hierarchy,..., The kth hierarchy are examples of the replication number in the present invention.

1-1-1. When the Number of Hierarchies is “2” FIG. 3 is an explanatory diagram of the storage charge range of each node Nn in each hierarchy when the number of hierarchies is “2”.

  FIG. 3A is an explanatory diagram of a storage charge range of each node Nn in the first hierarchy. Since the node ID and the storage charge range of each node Nn in the first hierarchy are the same as those in FIG. FIG. 3B is an explanatory diagram of a storage charge range of each node Nn in the second hierarchy. Each node Nn is assigned a node ID different from that of the first hierarchy, and a storage area different from that of the first hierarchy is defined.

In the example shown in FIG.
The node ID of the node N1 is “8000”, and the node N1 stores a replica having a file ID in the ID (8000 to bfff) area in the ID space.

  The node ID of the node N2 is “c000”, and the node N2 stores a replica having a file ID in the ID (c000 to ffff) area in the ID space.

  The node ID of the node N3 is “0000”, and the node N3 stores a replica having a file ID in the ID (0000-3fff) area in the ID space.

  The node ID of the node N4 is “4000”, and the node N4 stores a replica having a file ID in the ID (4000 to 7fff) area in the ID space.

  In the case of the example shown in the figure, since the ID spaces are evenly allocated by the four nodes Nn, the storage charge range in each node Nn is “1/4” and the number of layers is “2”. Therefore, the storage responsibility range of each node Nn satisfies the condition that it must be equal to or less than 1 / hierarchy (∵1 / 2 ≧ 1/4).

  The node ID of each node Nn in the first hierarchy and the second hierarchy can be determined by calculation.

The node ID of each node Nn in the first hierarchy is expressed by the following equation (1) using the node number of each node Nn itself, the total number of ID spaces, and the number of nodes Nn connected to the distributed storage system S. ) Can be determined according to. The “number of nodes Nn connected to the distributed storage system S” is an example of the number of node devices in the present invention, and will be referred to as “number of connected nodes” in the following description.
Node ID of the first layer = [node number-1] × [total number] / [number of connected nodes] (1)

  The node number is a number assigned to each node Nn connected to the distributed storage system S. The node number may be assigned in any way as long as it is a value that is greater than or equal to 1 and less than or equal to the number of connected nodes, and is assigned differently at each node Nn. For example, the node number of the node N1 is “1”, the node number of the node N2 is “2”, the node number of the node N3 is “3”, and the node number of the node N4 is “4”. When a node Nn is newly connected to the distributed storage system S, the node number of the new node Nn is set to “5”.

The node ID of each node Nn in the kth hierarchy can be determined according to the following equation (2) using the node ID of the first hierarchy of each node Nn itself, the total number of ID spaces, and the number of hierarchies. Note that k is a natural number of 2 or more.
Node ID of k-th layer = [Node ID of first layer] + ([Total number] / [Number of layers]) × (k−1) Expression (2)

  Note that the node ID may be a value that cannot be divided. In this case, the value that was not divisible could be rounded up or rounded down. Whether rounding up or down should be used consistently.

  FIG. 4A is an explanatory diagram of a storage charge range in the first hierarchy (FIG. 3A). The node number of each node Nn, the node ID of the first hierarchy determined according to the equation (1), and the storage charge range determined according to the node ID are associated with each other.

  FIG. 4B is an explanatory diagram of a storage charge range in the second hierarchy (FIG. 3B). The node IDs of the second hierarchy determined according to the equation (2) are associated with the storage charge ranges determined according to the node IDs.

  As described above, each node Nn stores a replica of a storage charge range determined in each hierarchy. For example, in the case of a replica having a file ID in the ID (0000-3fff) area, it is stored in the node N1 in the first hierarchy and is stored in the node N3 in the second hierarchy. In other words, the replica having the file ID in the ID (0000-3fff) area is stored redundantly in the node N1 and the node N3.

  In this manner, all replicas are configured to be stored in the nodes Nn in charge of storage determined in each hierarchy. When a new node Nn is connected to the distributed storage system S or when the node Nn leaves the distributed storage system S, each node Nn autonomously redetermines the node ID of each node Nn itself. Then, the assigned range of each node Nn is changed. Therefore, it is possible to quickly cope with the change of the storage charge range caused by the leaving node or the newly connected node. The node ID assigned to each node Nn in the first hierarchy is used as unique identification information for each node Nn. For example, the node ID in the first hierarchy is used as a routing key when various messages are transmitted / received between the nodes Nn.

  FIG. 5 is an explanatory diagram showing a state in which the new node N5 is connected to the distributed storage system S.

  The number obtained by adding 1 to the number of connected nodes before the new node Nn is connected to the distributed storage system S is determined as the node number of the new node Nn. In the case of the example shown in the figure, the number “5” obtained by adding 1 to the number of connected nodes “4” before the new node N5 is connected to the distributed storage system S is determined as the node number of the new node N5.

  First, the new node N5 transmits a participation message including the node number of the new node N5 itself to each of the nodes N1 to N4 connected to the distributed storage system S. Then, the new node N5 calculates the node ID in the first hierarchy and the node ID in the second hierarchy of the new node N5, respectively. The information on the total number and the number of hierarchies necessary for the calculation is inquired of each node N1 to N4 or management server to which the new node Nn is already connected, or is described in the catalog information, and the new node Nn is cataloged. What is necessary is just to acquire with reference to information.

  In the example shown in FIG. 5, the node ID of the first layer of the new node N5 is determined as “cccc” according to the above equation (1), and the node ID of the second layer is “4CCC” according to the above equation (2). It is determined.

  In addition, since each node N1 to N4 has been changed from “4” to “5” due to the addition of the new node N5, each of the nodes N1 to N4 is represented by the above equations (1) and (2). ), The node IDs of the respective layers of the nodes N1 to N4 themselves are determined again. In the example illustrated in FIG. 5, the node ID of the first hierarchy of the node N1 is “0000”, and the node ID of the second hierarchy is “8000”. The node ID of the first hierarchy of the node N2 is “3333”, and the node ID of the second hierarchy is “B333”. The node ID of the first hierarchy of the node N3 is “6666”, and the node ID of the second hierarchy is “E666”. The node ID of the first hierarchy of the node N4 is “9999”, and the node ID of the second hierarchy is “1999”.

  Then, each of the nodes N1 to N5 has a replica of the first hierarchy storage range determined based on the first hierarchy node ID, and a second hierarchy storage assignment determined based on the second hierarchy node ID. Each range replica is acquired and stored in its own storage unit described later.

  Then, after all the nodes N1 to N5 store the replicas of the storage charge range, if there are data files outside the storage charge range in the storage unit, the nodes N1 to N5 store the data files outside the storage charge range from the storage unit. delete.

  When the node Nn leaves the distributed storage system S, the leaving node Nn transmits a leaving message including the node number of the leaving node Nn itself to each node Nn. When each node Nn receives the leaving message, it updates the node number of each node Nn itself based on the node number of the leaving node Nn included in the leaving message. Depending on the node number of the leaving node Nn, the node number may not be updated. The update of the node number will be described in detail later.

  Further, since the number of connected nodes is changed by the leaving node Nn, each node Nn re-determines the node ID of each layer of each node Nn itself according to the above formulas (1) and (2). Then, the newly determined replica of the storage charge range is stored.

  Thus, even when a new node Nn is connected to the distributed storage system S or when the node Nn leaves the distributed storage system S, each node Nn autonomously The node ID is re-determined, and the storage charge range in each hierarchy is determined. That is, each node Nn itself performs storage control of the replica in charge of storage.

  Further, each node Nn is configured to delete replicas outside the storage charge range from the storage unit after all the nodes Nn have newly stored the storage charge range replicas by changing the storage charge range. This ensures that the data file is relocated to the node Nn in charge of storage without affecting the replica exchange being performed in the distributed storage system S when the storage charge range changes. Can do.

1-1-2. In the case of other number of hierarchies So far, the embodiment in which the number of hierarchies is “2” has been described. However, even if the number of hierarchies is “3” or more, according to the above formulas (1) and (2), The storage charge range of each node Nn can be determined. Here, an embodiment in which the number of hierarchies is “3” will be described.

  FIG. 6 is an explanatory diagram of a storage charge range of each node Nn in each tier when the number of tiers is “3”. FIG. 6 shows an example in which four nodes Nn of the nodes N1 to N4 are connected to the distributed storage system S. The number of connected nodes is “4”.

  In the example shown in the figure, the ID space is evenly allocated by the four nodes Nn, so that the storage charge range at each node Nn is “1/4” and the number of layers is “3”. Therefore, the storage responsibility range of each node Nn satisfies the condition that it must be equal to or less than 1 / hierarchy (∵1 / 3 ≧ 1/4).

  FIG. 6A is an explanatory diagram of a storage charge range of each node Nn in the first hierarchy. Since the node ID and the storage charge range of each node Nn in the first hierarchy are the same as those in FIG.

  FIG. 6B is an explanatory diagram of a storage charge range of each node Nn in the second hierarchy. FIG. 6C is an explanatory diagram of a storage charge range of each node Nn in the third hierarchy. According to the above equation (2), the node ID in the second hierarchy and the node ID in the third hierarchy of each of the nodes N1 to N4 are determined. In equation (2), k is “2” when determining the node ID in the second hierarchy, and k is “3” when determining the node ID in the third hierarchy. .

  In the example shown in FIG. 6B and FIG. 6C, the node ID of the second hierarchy of the node N1 is “5555”, and the node ID of the third hierarchy is “aaaa”. The node ID of the second hierarchy of the node N2 is “9555”, and the node ID of the third hierarchy is “eaaaa”. The node ID of the second hierarchy of the node N3 is “d555”, and the node ID of the third hierarchy is “2aaa”. The node ID of the second hierarchy of the node N4 is “1555”, and the node ID of the third hierarchy is “6aaa”.

  Each of the nodes N1 to N4 includes a replica of the storage charge range determined based on the node ID of the first hierarchy, a replica of the storage charge range determined based on the node ID of the second hierarchy, and the third hierarchy Each replica of the storage charge range determined based on the node ID is acquired and stored in the storage unit. For example, the replica of the file ID “3ccc” includes a storage node N1 defined in the first hierarchy, a storage node N4 defined in the second hierarchy, and a storage node defined in the third hierarchy. It is stored and saved in each of the three nodes N3.

  It is desirable that the storage charge ranges of the nodes Nn do not overlap in all layers. In a plurality of hierarchies, when the storage responsibility range of an arbitrary node Nn partially overlaps, when the node Nn leaves the distributed storage system S, the number of replicas on the distributed storage system S decreases by 2 or more. Because. If the storage charge range of an arbitrary node Nn in an arbitrary hierarchy is “total / number of hierarchies”, the storage charge ranges of all hierarchies are equal to the total number {(total / number of hierarchies) × hierarchy number = Total}. Therefore, if the storage responsibility range in each tier of the node Nn is equal to or less than “total / number of tiers”, the total storage responsibility range of each tier in a certain node device does not exceed the total number. When the total of the memory charge ranges exceeds the total number, the memory charge ranges are surely overlapped. However, since the sum of the memory charge ranges does not exceed the total number, it can be prevented that the memory charge ranges are surely overlapped.

1-2. Replica Input and Acquisition Replicas of data files that can be used in the acquisition distributed storage system S can be input from the management server, replica input server, or each node Nn. The input is to store a new replica in the node Nn connected to the distributed storage system S so that it can be used in the distributed storage system S. The management server manages the distributed storage system S. The replica insertion server manages the distribution of replicas. When the node Nn inputs a replica, only the node Nn that has received permission for replica input from the management server, the replica input server, or the like may be configured to input a data file. Hereinafter, a case where the node Nn inputs a replica will be described as an example.

  When a replica is inserted, first, the node Nn determines a replica file ID. As described above, as the file ID, a random value, a hash value, or the like that has the same bit length as the node ID and is distributed probabilistically uniformly in the ID space is used.

  Then, the node Nn that inputs the replica transmits the replica to the node Nn in charge of storage determined based on a predetermined rule. Based on the node number and the like of each node Nn connected to the distributed storage system S, the node Nn that inputs the replica can know the node Nn that is in charge of storage in each hierarchy of the replica that is input. The node number of each node Nn is inquired of the management server by the node Nn that inputs the replica or is described in the catalog information, and the node Nn that inputs the replica is obtained by referring to the catalog information. That's fine. Note that the node Nn to which the replica is inserted may not know the address information of all the nodes Nn of the distributed storage system S. However, each node Nn is based on the node ID by DHT routing described in JP 2006-197400 A and the like. By going through the node Nn, it is possible to make an inquiry to another node Nn.

  In this way, the node Nn that inputs the replica transmits the replica to the node Nn in charge of storing the replica in each hierarchy. The transmitted replica is received by each node Nn in charge of storage. The received replica is stored in a storage unit (to be described later) of each node Nn that has received the replica and is responsible for storing the replica. With the above operation, replica insertion is completed.

  Each node Nn can obtain a replica from another node Nn.

  A user node that is a node Nn that wants to obtain a replica refers to catalog information distributed to each node Nn from a management server, a replica server, or the like. Then, the user node acquires the file ID of the replica to be acquired from the description of the catalog information. The catalog information includes replica attribute information and the node number of each node Nn. Examples of the attribute information include a replica file ID, a data file evaluation value, a replica disclosure period, a data file name, a genre, an artist name, a performer, and a director name.

  Thereafter, the user node calculates the node ID of each node Nn based on the node number and the like of each node Nn. As described above, the node ID is different in each hierarchy. Which node ID is calculated may be determined by the user node, for example, by randomly selecting a hierarchy. Then, based on the file ID of the desired replica and the calculated node ID of each node Nn, a replica transmission request is made to the node Nn storing the replica. Note that the user node may not know the address information of all the nodes Nn of the distributed storage system S. In this case as well, each of the user nodes is based on the node ID by DHT routing described in JP-A-2006-197400. By going through the node Nn, it is possible to make an inquiry to another node Nn.

  When acquiring a plurality of replicas, the user node randomly selects a plurality of hierarchies, obtains the node IDs of the nodes in charge of storage Nn in the selected hierarchies, and distributes them from the plurality of nodes Nn. It is preferably configured to obtain a replica. This is because the load accompanying the acquisition of replicas when acquiring a large number of replicas can be distributed among the plurality of nodes Nn. In this case, the node ID in each hierarchy is used as unique identification information of each node Nn in each hierarchy. And node ID in each hierarchy is used as a routing key at the time of transmission / reception of various messages.

  In this way, each node Nn can input a replica to the distributed storage system S or acquire a replica desired to be acquired.

2. Configuration and Function of Node Nn Next, the configuration and function of the node Nn will be described with reference to FIG.

  FIG. 7 is a block diagram illustrating a schematic configuration example of the node Nn.

  Each node Nn includes a control unit 11, a storage unit 12, a buffer memory 13, a decoder unit 14, a video processing unit 15, a display unit 16, an audio processing unit 17, a speaker 18, and a communication unit 20. , And an input unit 21. The control unit 11, the storage unit 12, the buffer memory 13, the decoder unit 14, the communication unit 20, and the input unit 21 are connected to each other via a bus 22.

  The control unit 11 includes a CPU having a calculation function, a working RAM, a ROM that stores various data and programs, and the like. The control unit 11 functions as a computer of this embodiment. The storage unit 12 includes a hard disk (HD) that stores and stores (stores) replicas, various data, various programs, and the like. For example, the storage unit 12 stores a replica and a routing table in charge of storage. The buffer memory 13 temporarily stores received replicas and the like. The decoder unit 14 decodes (data expansion, decoding, etc.) encoded video data (video information), audio data (audio information), and the like. When the replica includes video and audio, the decoder unit 14 decodes the encoded video data and audio data included in the replica. The video processing unit 15 performs a predetermined drawing process on the decoded video data or the like and outputs it as a video signal. The display unit 16 is a CRT, a liquid crystal display, or the like that displays an image based on the video signal output from the image processing unit 15. The sound processing unit 17 performs D (Digital) / A (Analog) conversion of the decoded audio data into an analog audio signal, and then amplifies the signal by an amplifier and outputs the amplified signal. The communication unit 20 that outputs the audio signal output from the sound processing unit 17 as a sound wave performs communication control of information between other nodes Nn and the like through the network 8. The input unit 21 receives an instruction from the user and communicates an instruction signal corresponding to the instruction to the control unit 11. The input unit 21 is, for example, a keyboard, a mouse, or an operation panel. As the node Nn, a personal computer, an STB (Set Top Box), a communication karaoke apparatus, a TV receiver, or the like is applicable.

  The storage unit 12 stores address information of the node Nn itself.

  Further, after the node Nn is connected to the distributed storage system S, the storage unit 12 stores information on the node number of the node Nn itself, information on the number of connected nodes, and information on the number of layers.

  The storage unit 12 also stores information on the node ID of the node Nn itself and the file ID of the storage charge range.

  Further, the storage unit 12 stores catalog information and a management table. In the management table, information (title, file ID, release period, etc.) related to the replica stored in the storage unit 12 is registered.

  In such a configuration, the control unit 11 performs overall control by the CPU reading and executing a program stored in the storage unit 12 or the like.

  In addition, the control unit 11 executes the processing program of the present invention, whereby the first determination unit, the second determination unit, the storage control unit, the total number acquisition unit, the device number acquisition unit, the separation information reception unit, the separation in the present invention. It functions as message receiving means, node device number updating means, participation message receiving means, and the like. The processing program may be downloaded from a predetermined server on the network 8, for example, or may be recorded on a recording medium such as a CD-ROM and read via a drive of the recording medium. Anyway.

3. Processing in the node Nn at the processing node Nn will be described in detail with reference to the drawings.

3-1. Process Flow of Each Node Nn When New Node Nn is Connected FIG. 8 is a process flow when the new node Nn connects to the distributed storage system S. FIG. 8A shows processing of the control unit 11 in the new node Nn. FIG. 8B shows the processing of the control unit 11 in the node Nn that is already connected to the distributed storage system S.

  The processing from the new node Nn will be described with reference to FIG.

  First, the control unit 11 of the new node Nn sets the node number of the new node Nn itself (step S10). Specifically, the number obtained by adding 1 to the number of nodes Nn before the new node Nn connects to the distributed storage system S is set as the node number of the new node Nn itself. Information on the set node number is stored in the storage unit 12. Note that the information on the number of nodes Nn connected to the distributed storage system S is inquired to the node Nn to which the new node Nn is already connected, the management server, or the like. Alternatively, it may be described in the catalog information, and the new node Nn may be acquired by referring to the catalog information.

  Next, the control unit 11 functions as a participation message transmission unit, and transmits a participation message to all other nodes Nn connected to the distributed storage system S (step S11). The control unit 11 includes information on the node number of the new node Nn itself in the participation message.

  Next, the control unit 11 functions as a first determination unit and a second determination unit, and determines a storage charge range in each hierarchy of the new node Nn itself. First, the control unit 11 calculates and determines the node ID of the new node Nn itself in each hierarchy in accordance with Expression (1) and Expression (2) (Step S12). The control unit 11 stores the determination result in the storage unit 12.

  And the control part 11 determines the memory charge range in each hierarchy based on node ID of new node Nn itself in each determined hierarchy. And the control part 11 functions as a memory | storage control means, acquires the replica which has file ID of the memory charge range, and memorize | stores it in the memory | storage part 12 (step S13). If there is a replica that is already stored in the storage unit 12 among the replicas in the storage charge range, only the replica that is not stored is acquired from the other node Nn and stored in the storage unit 12. This is to prevent duplicate memory.

  When all the replicas in the storage charge range are stored in the storage unit 12, the control unit 11 displays an acquisition completion message indicating that the acquisition (storage) of the replica in charge of storage has been completed. To all the other nodes Nn connected to (step S14). For example, the acquisition completion message transmitted in step S14 is received in the process of step S23 described later.

Next, the control unit 11 receives an acquisition completion message from all the other nodes Nn connected to the distributed storage system S (step S15). At this time, the control unit 11 stands by until an acquisition completion message is received from all the other nodes Nn connected to the distributed storage system S.
The acquisition completion message received here is the acquisition completion message transmitted in step S22 of FIG. 8B executed by another node Nn described later. For example, when the control unit receives the acquisition completion message by the number obtained by subtracting “1” for its own node Nn from the number of connected nodes (number of connected nodes−1), the control unit receives the acquisition completion message from all other nodes Nn. What is necessary is just to determine as having done.

  When the control unit 11 receives the acquisition completion message from all the nodes Nn, if a replica outside the storage charge range of the node Nn is stored in the storage unit 12, the replica outside the storage charge range is stored. It deletes from the part 12 (step S16), and complete | finishes a process.

  Next, the processing of the control unit 11 in the node Nn connected to the distributed storage system S will be described with reference to FIG. The processing of the control unit 11 in the node Nn already connected to the distributed storage system S is started when the control unit 11 functions as a participation message receiving unit and receives a participation message from the new node Nn.

  First, the control unit 11 functions as a node device number updating unit and updates information on the number of connected nodes. Further, the control unit 11 re-determines the node ID of the node Nn itself in each hierarchy according to the above formulas (1) and (2) (step S20). The control unit 11 stores the updated number of connection nodes and the re-determined node ID in the storage unit 12.

  Next, the control unit 11 functions as a first determination unit and a second determination unit, and determines a storage charge range in each layer of the new node Nn itself based on the re-determined node ID. The control unit 11 functions as a storage control unit, acquires a replica having a file ID in the storage charge range from another node Nn, and stores it in the storage unit 12 (step S21). If there are replicas that are already stored in the storage unit 12 among the replicas in the storage area, the control unit 11 acquires only the unstored replicas from other nodes Nn in order to prevent duplicate storage. And stored in the storage unit 12.

  When the storage unit 12 stores all replicas in the storage charge range, the control unit 11 transmits an acquisition completion message to all other nodes Nn connected to the distributed storage system S (step S22). . The acquisition completion message transmitted at this time is received by the new node Nn in step S15. Note that the acquisition completion message transmitted in step S22 is also received in step S23 of the processing flow of FIG. 8B executed by each of the other nodes Nn.

  Next, the control unit 11 functions as an acquisition completion message receiving unit and receives an acquisition completion message from all the other nodes Nn connected to the distributed storage system S (step S23). At this time, the control unit 11 stands by until an acquisition completion message is received from all the other nodes Nn connected to the distributed storage system S. The acquisition completion message received here is the acquisition completion message transmitted in step S14 in FIG. 8A or step S22 in FIG. 8B executed by another node Nn. Whether or not the acquisition completion message has been received from all the other nodes Nn may be determined by a method similar to the method described in step S15.

  Then, when receiving the acquisition completion message from all the nodes Nn, the control unit 11 deletes the replica out of the storage charge range due to the change of the charge range from the storage unit 12 (step S25), and ends the process.

  If the replica in charge cannot be acquired from another node Nn in step S13 and step S21, the control unit 11 may acquire a replica from a replica input server that manages distribution of the replica.

  In step S10, the number obtained by adding “1” to the number of nodes Nn connected to the distributed storage system S is used as the node number of the new node Nn, but the node number of the new node Nn is set to “1”. Also good. When the node number of the new node Nn is “1”, each node Nn connected to the distributed storage system S updates the node number by incrementing the node number of each node Nn by “1”. Configure. In this case, the new node Nn notifies the node number of the set node Nn itself to the other node Nn, and each node Nn has a node number that does not overlap with other nodes Nn within the range of 1 or more and the number of connected nodes or less. Set. And each node Nn should just transmit the node number after setting to the other node Nn.

3-2. FIG. 9 is a processing flow of each node Nn when a leaving node is generated from the distributed storage system S. FIG. 9A shows the processing of the control unit 11 in the node Nn that has received the leaving information indicating that the node has left the leaving node. FIG. 9B shows processing of the control unit 11 in the node Nn that has received the leave message.

  With reference to FIG. 9 (A), the processing of the node Nn that has received the leaving information indicating that it has left the leaving node will be described.

  First, when receiving the leave information from the leaving node as the leave information receiving means, the control unit 11 functions as a leave message sending means and sends a leave message to all other nodes Nn connected to the distributed storage system S ( Step S30). The leaving message is a message for notifying that the leaving information has been received, and the control unit 11 includes the information on the node number of the leaving node in the leaving message. It is assumed that the node number of the leaving node is included in the leaving information.

  Next, the control unit 11 functions as a determination unit, and determines whether or not the node number of the node Nn stored in the storage unit 12 is larger than the node number of the leaving node (step S31). As a result of the determination, if the node number of the node Nn itself stored in the storage unit 12 is smaller than the node number of the leaving node (step S31: No), the process proceeds to step S33. On the other hand, when the node number of the node Nn stored in the storage unit 12 is larger than the node number of the leaving node as a result of the determination in step S41 (step S31: Yes), the control unit 11 It functions as an updating means, and updates the node number stored in the storage unit 12 by decrementing (subtracting) one (step S32).

  Subsequently, the control unit 11 functions as a node device number updating unit, updates the number of connected nodes, and re-determines the node ID of the node Nn itself in each hierarchy according to the above formulas (1) and (2) (steps). S33). Specifically, the number of connected nodes stored in the storage unit 12 is decremented (subtracted) by 1 and updated to determine the node ID again. The control unit 11 stores the updated number of connection nodes and the re-determined node ID in the storage unit 12.

  Next, the control unit 11 functions as a first determination unit and a second determination unit, and determines a storage charge range in each layer of the node Nn itself based on the re-determined node ID. The control unit 11 functions as a storage control unit, acquires a replica having a file ID in the storage charge range from another node Nn, and stores it in the storage unit 12 (step S34). If there are replicas that are already stored in the storage unit 12 among the replicas in the storage area, the control unit 11 acquires only the unstored replicas from other nodes Nn in order to prevent duplicate storage. And stored in the storage unit 12.

  When the storage unit 12 stores all replicas in the storage charge range, the control unit 11 transmits an acquisition completion message to all the other nodes Nn connected to the distributed storage system S (step S35). . The acquisition completion message transmitted at this time is received by the other node Nn in step S46 described later.

  Next, the control unit 11 functions as an acquisition completion message receiving unit, and receives acquisition completion messages from all the other nodes Nn connected to the distributed storage system S (step S36). At this time, the control unit 11 stands by until an acquisition completion message is received from all the other nodes Nn connected to the distributed storage system S. The acquisition completion message received here is the acquisition completion message transmitted in step S45 of FIG. 9B executed by another node Nn described later. Whether or not the acquisition completion message has been received from all the other nodes Nn may be determined by a method similar to the method described in step S15.

  Then, when receiving the acquisition completion message from all the nodes Nn, the control unit 11 deletes the replica out of the storage charge range due to the change of the charge range from the storage unit 12 (step S37), and ends the process.

  Next, the processing of the control unit 11 in the node Nn that has received the leave message will be described with reference to FIG. The process starts when the control unit 11 functions as a leaving message receiving unit and receives a leaving message.

  First, the control unit 11 functions as a determination unit, and determines whether or not the node number of the node Nn itself stored in the storage unit 12 is larger than the node number of the leaving node included in the leaving message (Step S11). S41). As a result of the determination, when the node number of the node Nn itself stored in the storage unit 12 is smaller than the node number of the leaving node (step S41: No), the process proceeds to step S43. On the other hand, when the node number of the node Nn stored in the storage unit 12 is larger than the node number of the leaving node as a result of the determination in step S41 (step S41: Yes), the control unit 11 It functions as an updating means and updates the node number stored in the storage unit 12 by decrementing (subtracting) one (step S42).

  Subsequently, the control unit 11 functions as a node device number updating unit, updates the number of connected nodes, and re-determines the node ID of the node Nn itself in each hierarchy according to the above formulas (1) and (2) (steps). S43). Specifically, the number of connected nodes stored in the storage unit 12 is decremented (subtracted) by 1 and updated to determine the node ID again. The control unit 11 stores the updated number of connection nodes and the re-determined node ID in the storage unit 12.

  Next, the control unit 11 functions as a first determination unit and a second determination unit, and determines a storage charge range in each layer of the node Nn itself based on the re-determined node ID. Then, the control unit 11 functions as a storage control unit, acquires a replica having a file ID in a storage charge range from another node Nn, and stores it in the storage unit 12 (step S44). If there are replicas that are already stored in the storage unit 12 among the replicas in the storage area, the control unit 11 acquires only the unstored replicas from other nodes Nn in order to prevent duplicate storage. And stored in the storage unit 12.

  When the storage unit 12 stores all the replicas in the storage charge range, the control unit 11 transmits an acquisition completion message to all the other nodes Nn connected to the distributed storage system S (step S45). .

  Next, the control unit 11 functions as an acquisition completion message receiving unit, and receives acquisition completion messages from all other nodes Nn connected to the distributed storage system S (step S46). At this time, the control unit 11 stands by until an acquisition completion message is received from all the other nodes Nn connected to the distributed storage system S. The acquisition completion message received here is the acquisition completion message transmitted in step S35 of FIG. 9A and step S45 of FIG. 9B executed by another node Nn. Whether or not the acquisition completion message has been received from all the other nodes Nn may be determined by a method similar to the method described in step S15.

  Then, when receiving the acquisition completion message from all the nodes Nn, the control unit 11 deletes the replica out of the storage charge range due to the change of the charge range from the storage unit 12 (step S47), and ends the process.

  In the above-described embodiment, the data file that is not in charge of storage is configured to be deleted. However, when the storage unit 12 has a margin, such as when the usage rate of the storage unit 12 is less than a predetermined value, You may memorize it without deleting.

  In the above-described embodiment, the node number information of each node Nn is transmitted and received between the nodes Nn in steps S11 and S30. However, these pieces of information may be described in the catalog information. In this case, each node Nn may refer to catalog information as necessary to obtain the node number of each node Nn.

  As described above, a plurality of data file replicas are generated, and each of the generated replicas is stored in a plurality of nodes Nn in duplicate. Each node Nn is configured to autonomously determine a plurality of storage areas defined in each hierarchy and store a replica of the determined storage area. Specifically, each node Nn stores a replica belonging to the storage charge range of the first hierarchy and a replica belonging to the storage charge range of the second hierarchy that is different from the storage charge range of the first hierarchy. It was configured as follows.

  Thereby, each node Nn itself can perform storage control of a replica in charge of storage without using a management server.

  Further, the storage area is determined based on the above formulas (1) and (2) according to the total number, the number of connected nodes, the number of hierarchies, and the like. As a result, even when a node Nn newly joins the distributed storage system S or when there is a leaving node, each node Nn autonomously re-determines the storage responsibility range, and each node Nn itself is responsible for storage It is possible to control storage of replicas.

  Further, in the present embodiment, the storage charge range of each node Nn is configured to be equal to or less than 1 / hierarchy, so that the storage charge range of each node Nn can be different in each hierarchy. Accordingly, each replica can always be stored in as many different nodes Nn as the number of layers. Therefore, even if any one of the nodes Nn leaves, the number of replicas is reduced by only 1. In other words, even when a maximum of (number of hierarchies-1) nodes leave at the same time, at least one replica can be guaranteed to remain in the distributed storage system S.

  In this embodiment, the storage charge range in each tier of each node Nn is configured to be equal to or less than 1 / hierarchical number, but is not limited to this. A storage charge range in a predetermined hierarchy among the number of hierarchies is determined, and a storage charge range for each layer other than the predetermined hierarchy may be determined differently for each storage charge range in the predetermined hierarchy. . With such a configuration, a storage charge range different from the predetermined hierarchy is determined at least in a hierarchy other than the predetermined hierarchy. Therefore, it is possible to prevent the ranges in charge of storage from overlapping at least between the predetermined hierarchy and a hierarchy other than the predetermined hierarchy. Further explanation regarding “at least a predetermined hierarchy and a storage charge range can be prevented” will be further described. If the storage charge range of each node Nn is larger than 1 / hierarchy, the predetermined hierarchy and the storage charge range. May overlap. For example, when the assigned range for one layer of a certain node Nn is two-thirds of the entire range, at least one-third of the assigned ranges of any other layer overlap. In order to completely prevent duplication, at least the storage charge range of each node Nn needs to be less than 1 / hierarchy.

  The distributed storage system S has been described on the assumption that it is formed by an algorithm using DHT, but the present invention is not limited to this.

4). Modified Example In the above-described embodiment, the storage responsibility range of each node Nn connected to the distributed storage system S is equally assigned to each node Nn in the ID space. In this case, the storage charge range of each node Nn may not be configured to be equal to or less than 1 / hierarchy.

  For example, it is determined according to the hard disk (HD) capacity of the storage unit 12 of each node Nn. For example, the node Nn having a larger HD capacity can be configured to increase the storage charge range so that the HD usage rate of each node Nn is uniform. In the case where the HD usage rate of each node Nn is configured to be uniform, for example, a technique described in detail in Japanese Patent Application No. 2009-123364 by the applicant of the present application may be used.

  Here, the determination method of the storage charge range of the second hierarchy when the storage charge range of each node Nn is not uniformly assigned in the first hierarchy will be described. In other words, this is a modified example of “how to shift the storage area”. Both are examples when two nodes Nn, nodes N1 and N2, are connected to the distributed storage system S.

  FIG. 10 is an explanatory diagram of a method for determining the storage charge range of the second hierarchy in the modified example.

  FIG. 10A is an explanatory diagram of a storage charge range of each node Nn in the first hierarchy, which is common to the determination methods 1, 2, and 3 in the second hierarchy. The storage charge range of the first tier may be determined according to the HD capacity of each node Nn as described above. In the example shown in FIG. 10A, the node ID of the node N1 is “0000”, and the node ID of the first layer of the node N2 is “a000”.

4-1. Method 1 for determining the storage area in the second hierarchy in the modified example 1
FIG. 10B is an explanatory diagram of a storage charge range of each node Nn in the second hierarchy in the case of the determination method 1. In accordance with the above equation (2), the node IDs in the second hierarchy of the nodes N1 and N2 are determined. In the example shown in FIG. 10B, the node ID of the node N1 is “8000”, and the node ID of the node N2 is “2000”.

  In the case of this configuration, when the node N2 whose storage charge range is less than or equal to the number of hierarchies leaves, the duplicate number of replicas is reduced by only one. On the other hand, when the node N1 whose storage charge range exceeds 1 / hierarchy leaves, the duplication number of replicas decreases by one or more. As described above, the duplication number of replicas may be reduced by one or more, but this is an effective configuration as a method for determining the storage responsibility range of a plurality of nodes Nn.

4-2. Method 2 for determining storage range in second layer in modified example 2
FIG. 10C is an explanatory diagram of a storage charge range of each node Nn in the second hierarchy in the case of the determination method 2. According to the following formula (3), node IDs in the second hierarchy of the nodes N1 and N2 are determined. Note that k is a natural number of 2 or more, and k is 2 here.
Node ID of the k-th layer = [Node ID of the first layer] + [Storage range of the node Nn having the largest storage range] × (k−1) Expression (3)

  Since the node Nn having the largest storage charge range is the node N1, the node ID of each node Nn in the second hierarchy is determined as a value shifted by the storage charge range of the node N1. In the example shown in FIG. 10C, the node ID of the node N1 is “4000”, and the node ID of the node N2 is “a000”.

  In the case of this configuration, since there is a node Nn (node N1) in which the storage charge range exceeds 1 / hierarchy, the replica duplication number decreases by one or more when any node Nn leaves. there is a possibility. However, this is an effective configuration as a method for determining the storage responsibility range of a plurality of nodes Nn.

4-3. Method 3 for determining storage range in second layer in modified example 3
FIG. 10D is an explanatory diagram of a storage charge range of each node Nn in the second hierarchy in the case of the determination method 3. According to the following formula (4), node IDs in the second hierarchy of the nodes N1 and N2 are determined. Note that k is a natural number of 2 or more, and k is 2 here.
Node ID of the k-th layer = [Node ID of the first layer] + [Storage range of the node Nn with the smallest storage range] × (k−1) Expression (4)

  Since the node Nn having the smallest storage charge range is the node N1, the node ID of each node Nn in the second hierarchy is determined as a value shifted by the storage charge range of the node N1. In the example shown in FIG. 10D, the node ID of the node N1 is “6000”, and the node ID of the node N2 is “0000”.

  In the case of this configuration, there is a possibility that the number of replicas may be reduced by one or more when any node Nn leaves. However, this is an effective configuration as a method for determining the storage responsibility range of a plurality of nodes Nn.

3 IX
4a, 4b ISP
5a, 5b DSL line company 6 FTTH line company 7 Communication line 8 Network 9 Overlay network (P2P network)
11 Control Unit 12 Storage Unit 13 Buffer Memory 14 Decoder Unit 15 Video Processing Unit 16 Display Unit 17 Audio Processing Unit 18 Speaker 20 Communication Unit 21 Input Unit 22 Bus Nn Node

Claims (6)

A data file that can be transmitted and received between a plurality of node devices that can communicate with each other via a network is replicated as a replica to which a replication number corresponding to a predetermined replication number is assigned,
The data file corresponding to the replicated replica can be identified, and for a plurality of identification information to be assigned to the data file,
In order to distribute and store the plurality of identification information in the plurality of node devices, a part of the plurality of identification information at a predetermined replication number in the predetermined number of replicas is determined as a storage charge range. A determination means;
A second determination for determining a storage charge range for each copy number different from the copy number determined by the first determination means, for each of the copy numbers determined by at least the storage charge range determined by the first determination means; Means,
Storage control means for storing in the storage means a replica of the storage charge range determined by the first determination means and the second determination means;
Total number acquisition means for acquiring the total number of the plurality of identification information;
Device number acquisition means for acquiring the number of node devices indicating the number of the plurality of node devices;
The leaving information receiving means for receiving the leaving node device's leaving information from the leaving node device leaving the distributed storage system constituted by the plurality of node devices, or the leaving node receiving the leaving information from the other node device A leave message receiving means for receiving a leave message of the node device;
When the leaving information receiving means is provided, when the leaving information is received, or when the leaving message receiving means is provided, when the leaving message is received, the leaving information or the leaving message is added. Node device number updating means for updating the number of node devices based on;
With
The first determination unit is configured to equalize the storage range for each of the plurality of node devices at the predetermined replication number based on the number of node devices updated by the node device number updating unit and the total number. Redetermined,
The second determining unit re-sequentially stores the storage responsibility ranges of the plurality of node devices for each different replication number based on the number of node devices updated by the node device number updating unit and the total number. Decide
The storage control unit causes the storage unit to store a replica of the storage charge range re-determined by the first determination unit and the second determination unit . A data file that can be transmitted and received between a plurality of node devices that can communicate with each other via a network is replicated as a replica to which a replication number corresponding to a predetermined replication number is assigned,
The data file corresponding to the replicated replica can be identified, and for a plurality of identification information to be assigned to the data file,
In order to distribute and store the plurality of identification information in the plurality of node devices, a part of the plurality of identification information at a predetermined replication number in the predetermined number of replicas is determined as a storage charge range. A determination means;
A second determination for determining a storage charge range for each copy number different from the copy number determined by the first determination means, for each of the copy numbers determined by at least the storage charge range determined by the first determination means; Means,
Storage control means for storing in the storage means a replica of the storage charge range determined by the first determination means and the second determination means;
Total number acquisition means for acquiring the total number of the plurality of identification information;
Device number acquisition means for acquiring the number of node devices indicating the number of the plurality of node devices;
Participation message receiving means for receiving a participation message from a new node device newly participating in the distributed storage system including the plurality of node devices;
Node device number updating means for updating the number of node devices based on the participation message when the participation message is received by the participation message receiving means;
With
The first determination unit is configured to equalize the storage range for each of the plurality of node devices at the predetermined replication number based on the number of node devices updated by the node device number updating unit and the total number. Redetermined,
The second determining unit re-sequentially stores the storage responsibility ranges of the plurality of node devices for each different replication number based on the number of node devices updated by the node device number updating unit and the total number. Decide
The storage control unit causes the storage unit to store a replica of the storage charge range re-determined by the first determination unit and the second determination unit . A data file that can be transmitted and received between a plurality of node devices that can communicate with each other via a network is replicated as a replica to which a replication number corresponding to a predetermined replication number is assigned,
The data file corresponding to the replicated replica can be identified, and for a plurality of identification information to be assigned to the data file,
In order to distribute and store the plurality of identification information in the plurality of node devices, a part of the plurality of identification information at a predetermined replication number in the predetermined number of replicas is determined as a storage charge range. A decision step;
A second determination for determining a storage charge range for each copy number different from the copy number determined in the first determination step for each of the copy numbers at least for the storage charge range determined in the first determination step; Steps,
A storage control step for storing in the storage means a replica of the storage charge range determined by the first determination step and the second determination step;
A total number obtaining step for obtaining a total number of the plurality of pieces of identification information;
A device number acquisition step of acquiring a node device number indicating the number of the plurality of node devices;
A leaving information receiving step for receiving leaving information of the leaving node device from a leaving node device that leaves the distributed storage system configured by the plurality of node devices, or the leaving from another node device that has received the leaving information A leave message receiving step for receiving a leave message of the node device;
When the leaving information is received by the leaving information receiving step, or when the leaving message is received by the leaving message receiving step, the number of node devices is updated based on the leaving information or the leaving message. Updating the number of node devices to perform,
A program for causing a computer to execute
In the first determination step, based on the number of node devices updated by the node device number updating step and the total number, the storage charge ranges for the plurality of node devices at the predetermined replication number are made equal. Redetermined,
Based on the number of node devices updated by the node device number updating step and the total number, the second determining step is repeated so that storage ranges of the plurality of node devices are equalized for each of the different replication numbers. Decide
The storage control step causes a storage unit to store a replica of a storage charge range re-determined by the first determination step and the second determination step . A data file that can be transmitted and received between a plurality of node devices that can communicate with each other via a network is replicated as a replica to which a replication number corresponding to a predetermined replication number is assigned,
The data file corresponding to the replicated replica can be identified, and for a plurality of identification information to be assigned to the data file,
In order to distribute and store the plurality of identification information in the plurality of node devices, a part of the plurality of identification information at a predetermined replication number in the predetermined number of replicas is determined as a storage charge range. A decision step;
A second determination for determining a storage charge range for each copy number different from the copy number determined in the first determination step for each of the copy numbers at least for the storage charge range determined in the first determination step; Steps,
A storage control step for storing in the storage means a replica of the storage charge range determined by the first determination step and the second determination step;
A total number obtaining step for obtaining a total number of the plurality of pieces of identification information;
A device number acquisition step of acquiring a node device number indicating the number of the plurality of node devices;
A participation message receiving step of receiving a participation message from a new node device newly participating in the distributed storage system including the plurality of node devices;
A node device number updating step of updating the number of node devices based on the participation message when the participation message is received by the participation message receiving step;
A program for causing a computer to execute
In the first determination step, based on the number of node devices updated by the node device number updating step and the total number, the storage charge ranges for the plurality of node devices at the predetermined replication number are made equal. Redetermined,
Based on the number of node devices updated by the node device number updating step and the total number, the second determining step is repeated so that storage ranges of the plurality of node devices are equalized for each of the different replication numbers. Decide
The storage control step causes a storage unit to store a replica of a storage charge range re-determined by the first determination step and the second determination step . A data file that can be transmitted and received between a plurality of node devices that can communicate with each other via a network is replicated as a replica to which a replication number corresponding to a predetermined replication number is assigned,
The data file corresponding to the replicated replica can be identified, and for a plurality of identification information to be assigned to the data file,
In order for the node device to store the plurality of identification information in the plurality of node devices in a distributed manner, a part in charge of storing a part of the plurality of identification information at a predetermined replication number in the predetermined number of replicas A first determining step to determine as
The node device makes a storage charge range for each copy number different from the copy number determined in the first determination step at least for each copy number for the storage charge range determined by the first determination means. A second determining step for determining;
A storage control step in which the node device stores, in a storage unit, a replica of the storage charge range determined by the first determination step and the second determination step;
A total number obtaining step for obtaining a total number of the plurality of pieces of identification information;
A device number acquisition step of acquiring a node device number indicating the number of the plurality of node devices;
A leaving information receiving step for receiving leaving information of the leaving node device from a leaving node device that leaves the distributed storage system configured by the plurality of node devices, or the leaving from another node device that has received the leaving information A leave message receiving step for receiving a leave message of the node device;
When the leaving information is received by the leaving information receiving step, or when the leaving message is received by the leaving message receiving step, the number of node devices is updated based on the leaving information or the leaving message. Updating the number of node devices to perform,
Including
In the first determination step, based on the number of node devices updated by the node device number updating step and the total number, the storage charge ranges for the plurality of node devices at the predetermined replication number are made equal. Redetermined,
Based on the number of node devices updated by the node device number updating step and the total number, the second determining step is repeated so that storage ranges of the plurality of node devices are equalized for each of the different replication numbers. Decide
The distributed storage method, wherein the storage control step stores, in a storage unit, a replica of the storage charge range re-determined by the first determination step and the second determination step . A data file that can be transmitted and received between a plurality of node devices that can communicate with each other via a network is replicated as a replica to which a replication number corresponding to a predetermined replication number is assigned,
The data file corresponding to the replicated replica can be identified, and for a plurality of identification information to be assigned to the data file,
In order for the node device to store the plurality of identification information in the plurality of node devices in a distributed manner, a part in charge of storing a part of the plurality of identification information at a predetermined replication number in the predetermined number of replicas A first determining step to determine as
The node device makes a storage charge range for each copy number different from the copy number determined in the first determination step at least for each copy number for the storage charge range determined by the first determination means. A second determining step for determining;
A storage control step in which the node device stores, in a storage unit, a replica of the storage charge range determined by the first determination step and the second determination step;
A total number obtaining step for obtaining a total number of the plurality of pieces of identification information;
A device number acquisition step of acquiring a node device number indicating the number of the plurality of node devices;
A participation message receiving step of receiving a participation message from a new node device newly participating in the distributed storage system including the plurality of node devices;
A node device number updating step of updating the number of node devices based on the participation message when the participation message is received by the participation message receiving step;
Including
In the first determination step, based on the number of node devices updated by the node device number updating step and the total number, the storage charge ranges for the plurality of node devices at the predetermined replication number are made equal. Redetermined,
Based on the number of node devices updated by the node device number updating step and the total number, the second determining step is repeated so that storage ranges of the plurality of node devices are equalized for each of the different replication numbers. Decide
The distributed storage method, wherein the storage control step stores, in a storage unit, a replica of the storage charge range re-determined by the first determination step and the second determination step .

Images