JP2008234536A - Node device, transfer node determination method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a DHT node for more flexibly constructing a distributed hash table (DHT). <P>SOLUTION: The DHT node 3 has a finger table 35 in which a transfer candidate DHT node is registered when a message to be received by a DHT node shared by a specific hash value is transmitted without specifying a receiving destination node. A management part 36 selects the number of DHT nodes which should be registered in the finger table 35 from among the DHT nodes except the own-node to be fixed according to a specified parameter. In that case, the management part 36 selects a node device to be registered in the finger table based on a hash value obtained by adding r-th power (r is a real number selected from a range of 0≤r<m, m is the number of bits of a hash space managed by the DHT) of 2 to a hash value regarding the own-node and a hash value regarding the DHT nodes except the own-node. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a node device, a forwarding node determination method, and a program used for a distributed hash table.

  The Chord method (Non-patent Document 1) is known as one of the distributed hash table (DHT) methods.

In Chord, a method for DHT construction is fixed. Specifically, the short-circuit path design method is equal to the modified binary tree search, and the cost for the search is O (log ₂ N) when the number of DHT nodes is N. For this reason, it has been difficult to flexibly meet the demands of various applications. In particular, there is a trade-off relationship between the communication cost for maintaining the function and the time required for the inquiry response or the number of recursive stages, and this cannot be controlled.
I. Stoica, R. Morris, D. Karger, MF Kaashoek, and H. Balakrishnan. Chord: A scalable peer-to-peer lookup service for internet applications.In Proceedings of ACM SIGCOMM, August 2001.

  In Chord known as one of DHT, for example, it is not possible to flexibly construct a DHT according to the needs of applications using the DHT.

  The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a DHT node apparatus, a forwarding node determination method, and a program that can construct a DHT more flexibly.

  The present invention relates to a message that should be received by a node device that shares a specific hash value in one node device that constitutes a distributed hash table system that includes a plurality of node devices that share different hash value ranges. A finger table in which information of a plurality of node devices that are candidates for transferring the message is to be transmitted when the device cannot be specified, and all node devices other than the own node constituting the distributed hash table system A selection unit that selects a number of node devices to be registered in the finger table according to a designated parameter, and each node device constituting the distributed hash table system has a different hash value. Are associated with each other, and the selecting means associates with the own node. 2 is added to the hash value (r is a real number selected from the range of 0 ≦ r <m, and m is the number of bits in the hash space managed by the distributed hash table system). A node device to be registered in the finger table is selected based on the obtained hash value and a hash value associated with the node device other than the own node.

  According to the present invention, a DHT can be constructed more flexibly.

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

  FIG. 1 shows a configuration example of a distributed hash table (DHT (Distributed Hash Table)) system according to an embodiment of the present invention.

  In FIG. 1, 1 is a client, 3 is a DHT node, and 8 is a network.

  A plurality of DHT nodes 3 cooperate to form one distributed hash table. In FIG. 1, eight DHT nodes are shown as the DHT nodes constituting the distributed hash table, but this is an example, and the number of DHT nodes constituting the distributed hash table system is arbitrary. Further, the number of DHT servers can vary depending on participation and withdrawal.

  A plurality of DHT nodes 3 constituting one distributed hash table can communicate with each other via the network 8. The network 8 may be any network.

  The client 1 can access any DHT node 3 via the network 8. In FIG. 1, only one client computer is shown. Of course, there may be a plurality of client computers. A configuration in which the client 1 is locally connected to any DHT node 3 is also possible.

  Now, in the distributed hash table of the present embodiment, a case is considered in which a hash space is formed as a closed number line in which both ends are connected in a hash space composed of a set of hash values. The plurality of DHT nodes 3 are associated with different hash values in the hash space. In this case, each DHT node 3 is adjacent to two other DHT nodes 3.

  Here, a hash value corresponding to a certain DHT node is defined as α, and a DHT node associated with a hash value closest to α in the direction in which the hash value increases in the hash space is determined with respect to the certain DHT node. A DHT node associated with a hash value closest to α in a direction in which the hash value decreases is referred to as an “increasing direction side adjacent node” and is referred to as a “decreasing direction side adjacent node” for the certain DHT node Shall.

  In the present embodiment, assuming that a hash value corresponding to a certain DHT node is x1, and a hash value corresponding to a decreasing direction adjacent node of the certain DHT node is x0, the certain DHT node is x0 <h ≦ A case where a sharing method for sharing the hash value h in the range of x1 is adopted will be described as an example. Of course, other sharing methods are possible.

  For example, the DHT node A1 that shares the hash value h in the range of h0 <h ≦ h1 has an identifier such that the hash value h of key information (hereinafter referred to as an identifier) is in the range of h0 <h ≦ h1. Then, a database in which the identifier and data searched using the identifier as a key are stored and managed.

  Further, the DHT node A2, which is an adjacent node on the increasing direction side of the DHT node A1, has an identifier for which the hash value h of the identifier becomes h1 <h ≦ h2, where the hash value corresponding to this is x2. And a database in which data searched using the identifier as a key is stored and managed.

  The DHT node A1 has a hash value corresponding to the decreasing direction side adjacent node A0 (or information that is the source), in addition to the hash value (or information that is the source) of the own node, If the hash value (or information that is the basis) corresponding to the increasing direction side adjacent node A2 is stored, the allocation range of the own node A1 and the allocation range of the increasing direction side adjacent node A2 are specified. (In this case, the sharing range of the decreasing direction side adjacent node A0 is not known).

As the hash function, it is preferable to use a function that has a sufficiently long number of bits and is uniformly assigned. For example, SHA-1 or MD5 can be adopted as such a hash function. The hash space is represented by an m-bit integer of 0 to 2 ^m −1 (m = 160 for SHA-1 and m = 128 for MD5).

  As a method for determining a hash value to be associated with each DHT node 3, for example, a method of using a hash value obtained by applying a hash function to an address (for example, an IP address) of the DHT node 3 may be adopted. Alternatively, other methods may be adopted.

  FIG. 2 shows a configuration example of the DHT node according to the present embodiment.

  As shown in FIG. 2, the DHT node 3 includes a message processing unit 31, a local node information storage unit 32, an adjacent node information storage unit 33, a database 34, a finger table 35, a finger table management unit 36, and a communication unit 37. ing. Further, the finger table management unit 36 includes a parameter setting unit 361, a calculation unit 362, and a node search unit 363.

  The communication unit 37 is for performing message communication with the client 1 and other DHT nodes 3 via the network 8. When the client 1 is connected locally to the DHT node 3, an interface for connecting to the client 1 may be provided.

  When the message processing unit 31 receives an inquiry message from the client 1 or another DHT node 3 via the communication unit 37, the message processing unit 31 processes the received inquiry message and, if necessary, the client via the communication unit 37 1 for transmitting a response message to 1 or transferring an inquiry message to another DHT node 3.

  The own node information storage unit 32 stores information about the own node such as an address of the own node and a hash value corresponding to the own node.

  The adjacent node information storage unit 33 stores a hash value corresponding to the adjacent node in the decreasing direction with respect to the own node (it may also store the address thereof). Further, the address and the hash value corresponding to the address are stored for the increasing direction side adjacent node.

  The database 34 stores, for data having an identifier that gives a hash value belonging to the hash space shared by the own node, the identifier and data searched using the identifier as a key. FIG. 3 shows an example of the database 34. The data stored in the database 34 may be data that the client 1 wants to finally acquire, or may be information such as a URL for acquiring data that the client 1 wants to finally acquire. it can.

The finger table (route table) 35 corresponds to the address of the DHT node and the DHT node for several DHT nodes (M = f _amp × m DHT nodes described later) that are candidates for transferring the inquiry message. The hash value to be stored is stored. FIG. 4 shows an example of the finger table 35. 4 is an entry index (corresponding to “next” in FIG. 5).

  The finger table management unit 36 is for managing the finger table 35, as will be described in detail later.

  When the hash value of the address of a certain DHT node is used as the hash value corresponding to the DHT node, only the address is stored in the own node information storage unit 32, the adjacent node information storage unit 33, and the finger table 35. Alternatively, the hash value may be calculated each time without being stored.

  In the DHT node 3 of the present embodiment, management of information related to the own node, participation / leaving control, management of information related to adjacent nodes, management of the database, and DHT responsible for an arbitrary hash value in the hash space to which the own node belongs. Basically, a conventional method may be used for the node search or the like.

  Next, a search for a DHT node registered in the finger table 35 will be described.

  FIG. 5 shows an example of an algorithm (an algorithm for determining a Finger route) in which the finger table management unit 36 of the DHT node 3 according to this embodiment searches for a DHT node registered in the finger table 35.

It is assumed that parameter m and parameter f _amp (Finger Amplification Ratio) (a real value greater than 0) are set in advance in parameter setting unit 361. M is the bit width of the hash space (0 to 2 ^m −1), and is a value determined according to the hash function employed (in the case of SHA-1, m = 160, in the case of MD5) M = 128). Further, f _amp can be arbitrarily set. For example, the user may be able to set it appropriately, or another application may be specified.

The number of DHT nodes registered in the finger table 35 (Finger's routing table entry number) is M as described above (that is, it can be controlled by adjusting f _amp ).

  Next used in the algorithm is an integer that varies from 1 to M, and is a parameter (Finger's routing table entry index) that represents the order of DHT nodes registered in the finger table 35.

In the above algorithm, first, the calculation unit 362 sets 0 to next (step S1) and sets f _amp × m to M (step S2).

However, M is prepared so as to be an integer value close to m · f _amp . At that time, various approximation methods such as rounding off after the decimal point, rounding up, rounding down, and other rounding methods in units of two or more can be adopted.

  Note that step S1 and step S2 may be executed in the reverse order of the above, or may be executed simultaneously.

  Thereafter, the process is repeated while incrementing next by one. That is, first, next = 1 is set (step S3), the processing of steps S6 to S8 is performed, and after waiting for a specified number of seconds (step S9), next is incremented by 1 (step S3), and the same processing is repeated. When next exceeds M (step 4), next is returned to 1 again (step S5), and the same processing is repeated.

In the iterative process, the calculation unit 362 sets m × (next−1) / M to r (step S6), and sets n + 2 ^r to the target value target on the DHT hash space (step S7). Then, the node search unit 363 executes finger [next] = find_succ (target) (step S8).

  Note that n is a hash value corresponding to the own node.

For example, in the case of f _amp = 2.0 and m = 160, when next = 1, finger [1] = find_succ (n + 1) is executed from r = 0 and target = n + 1.

When next = 2, finger [1] = find_succ (n + 2 ^1/2 ) is executed from r = ^1/2 and target = n + 2 ^1/2 .

When next = 319, finger [1] = find_succ (n + 2 ^318/2 ) is executed from r = ^318/2 and target = n + 2 ^318/2 .

When next = 320, finger [1] = find_succ (n + 2 ^319/2 ) is executed from r = ^319/2 and target = n + 2 ^319/2 .

  This finger [next] = find_succ (target) searches for the DHT node closest to the hash value represented by the target (in this case, a hash value successor), and uses information about the DHT node as the finger table 35. To fill in the next entry.

  If the evaluation of the proximity between the hash value represented by the target and the position of each DHT node in the hash space does not contradict the algorithm of message transfer (Key Based Routing), the position of the node in the hash space The absolute value of the difference between them may be used, or a restriction may be given based on the order of the hash space.

  In the above, as in the case of Chord, a case where the hash value successor (the closest DHT node in the hash space increasing direction in the hash space) has corresponding data is shown as an example, but the algorithm details include various variations. There can be. For example, the same can be realized when the predecessor (the DHT node closest in the hash value phenomenon direction on the hash space) has corresponding data.

The problem of Chord is that the algorithm lacks flexibility. In this embodiment, a mechanism for changing the method for setting the Finger according to the request from the application is introduced in the module for setting the Finger in the DHT node. be able to. More specifically, the granularity of selection of a finger destination can be controlled by adjusting a parameter f _amp that is arbitrarily determined based on, for example, DHT characteristics required by an application.

FIGS. 6 to 8 show the state of the target obtained when f _amp = 1.0, 2.0, and 3.0, respectively. 6 to 8, the circle represents the hash space when the hash space is formed in a loop shape assuming that both ends of the hash space are connected, and the top of the circle (the direction at 0 o'clock) is the self-space. Indicates the position of the hash value corresponding to the node, and the calculation result of how the hash value of the target for the node is arranged at the position where the tip of each arrow coming out of the node intersects the circle (The arrangement of the target corresponding to less than 1/16 of the circumference is omitted). FIG. 6 is the same as the conventional Chord.

As is apparent from FIGS. 6 to 8, the distribution of the target becomes denser as _famp is increased, and more areas on the circumference can be reached with a shorter number of hops. I understand.

In other words, when f _amp = 1.0, the same algorithm as Chord is obtained, and when f _amp <1.0, the time required for the query and the average number of recursive stages increase compared to Chord, but the maintenance cost can be reduced. When f _amp > 1.0, the maintenance cost is higher than that of Chord, but it can be expected that the time required for the inquiry (the average number of recursive stages) can be reduced.

  Note that the algorithm of FIG. 5 is an example, and other algorithms can be implemented.

  Next, message transfer in DHT will be described.

  As described above, the case where the hash value successor has the corresponding data is shown as an example. However, there may be various variations such as the case where the predecessor has the corresponding data.

  FIG. 9 shows an example of a processing procedure when an inquiry message is received by the DHT node according to the present embodiment.

  When the client 1 wants to acquire data corresponding to a desired identifier from the distributed hash table system, the client 1 transmits an inquiry message including the desired identifier to any DHT node 3. When the client 1 knows the addresses of a plurality of DHT nodes 3, the client 1 may send an inquiry message to any of the DHT nodes 3.

  When the DHT node 3 receives an inquiry message from the client 1 (or receives an inquiry message transmitted by the client 1 and forwarded by one or more other DHT nodes) (step S11), first, the DHT node 3 receives the inquiry message. The hash value (destination hash value) hm of the included identifier is calculated (step S12).

  Then, it is checked whether or not this is included in the allocation range of the own node (step S13). If it is included in the allocation range of the own node, the database is searched in the own node (step S14), data corresponding to the identifier is acquired, and a response message including the data is returned to the client 1. (Step S15).

Next, when it is not included in the shared range of the own node, it is checked whether it is included in the shared range of the adjacent node on the increasing direction side (step S16). If it is included in the sharing range of the increasing direction side adjacent node, the inquiry message is transferred to the increasing direction side adjacent node (step S17). Here, if the hash value hm exists between the hash value n corresponding to the own node and the n successors, the inquiry message is transferred to the n successors.
In other cases, the transfer destination node is selected with reference to the finger table 35 (step S18), and the inquiry message is transferred to the selected transfer destination node (step S19).

  Here, the hash value hm is compared with the target [i] for each entry i (= next) in the finger table 35, and the target [i] is forward in the hash space (the hash value increases compared to hm). Direction) to find the maximum i (that is, the maximum i in the range from n to hm by following the hash space in the forward direction) on the own node side (the side on which the hash value decreases) The message is transferred to the DHT node entered in the entry i of the finger table 35 corresponding to i.

  When the above algorithm is used, the distance between the target [i] and the position of each DHT node in the hash space exists in the space from n to target [i] when the hash space is considered in the forward direction. It is preferable to calculate the distance to target [i] from the DHT nodes to be performed (constrained in the forward direction of the hash space).

  Note that the algorithm of FIG. 9 is an example, and other algorithms can be implemented. For example, in the algorithm of FIG. 9, the DHT node that holds the data related to the identifier included in the inquiry message returns a response message to the client, but first passes through the DHT node that received the inquiry message from the client. The response message may be returned to the client. Further, the format of the inquiry message between the client and the DHT node may be different from the format of the inquiry message between the DHT nodes.

  Next, a case will be described in which the present distributed hash table system is used to construct an individual product distribution management / traceability system using an electronic tag.

  In general, an ID assigned to each individual product is composed of three layers of company ID, product type ID, and individual product ID.

  FIG. 10 shows GID-96 (EPCglobal. EPC tag data standards version 1.3. EPCglobal Ratified Specification, March 2006) in EPCglobal as an example of IDs assigned to individual products.

It is conceivable to use DHT in order to manage a large amount of information of individual items tagged with such IDs inexpensively and scalable. However, since the service quality required for the traceability system and the cost to be applied are different for each product, it is difficult to cover all of them with a single DHT. Therefore, for example, a DHT is assigned for each company and each product type, or a DHT is assigned for each f _amp , and an ID including an individual product ID and information corresponding thereto are managed.

  In this case, each DHT is a service provided to each customer by each company, and the company wants to determine the trade-off between quality and cost. However, when using the conventional Chord, one of the most proven algorithms in DHT, this trade-off cannot be controlled mainly because the Finger algorithm is fixed.

Therefore, by implementing the DHT system described so far in each DHT in the traceability system, this trade-off can be controlled. Specifically, the application sets f _amp obtained from the request, and constructs a finger table according to the algorithm of this embodiment. Thereby, trade-off can be controlled. For example, in the case of using a traceability system, in general, it means that when an accident occurs, general food information is usually set so that it can be maintained at a low cost (f _amp = 1.0). It is conceivable to manage with DHT. On the other hand, when credit through traceability is useful for improving product value, such as branded foods, the communication cost is high to improve the reaction speed of the system, but the inquiry can be answered in a short time (f _amp = about 4.0) DHT may be set, and information management may be performed on the DHT.

By the way, in the flowchart of FIG. 5, target = n + 2 ^r , but there are several variations in the calculation method of the portion corresponding to 2 ^r .

For example, for some operations, the target = n + ^{b r,} may control both the b and r. Alternatively, by using the integer number i of 0 or more suitable range instead of r, it may be target = n + ^{b i.} In this case, the appropriate range is a range where b ⁱ or b ^r does not exceed 2 ^m (where m is the bit width of the hash function used in Chord), that is, a range that does not go around the hash space. For example, if the formula of target = n + b ⁱ is used and m = 160, if b = 1.5, i is up to 437.

In general, calculation of floating point numbers in a computer is expensive. If it is desired to avoid this cost, the parameters used in the calculation of the target may be limited to combinations that can be calculated using a fixed-point or binary fraction. A combination of 2 ^r , b ⁱ , or b ^r may be prepared, and the user may select a combination that suits the purpose from the combination.

Furthermore, a table corresponding to 2 ^r, b ^i, or b ^r necessary for target calculation may be prepared in advance and looked up as necessary. In this case, it is necessary to mount a memory for this purpose, but the calculation can be performed at low cost.

  Below, the case where a parameter is dynamically adapted is demonstrated.

Actually, the demand for a certain product fluctuates, and even if the amount of traffic required to maintain the Chord network increases during periods when demand is high, reducing the number of recursive stages for inquiries will reduce the total amount of traffic. The case where it is possible is considered. In preparation for such a case, each node of the DHT may introduce a mechanism for dynamically changing a parameter (for example, f _amp ) that determines a trade-off.

However, in DHT, communication is always local and it is difficult to know how the entire traffic is generated. Therefore, each node determines whether to increase or decrease f _amp from the frequency of inquiries arriving at itself and the estimated current network density.

From the density of the network, the number of finger tables that will be actually used when f _amp is changed can be estimated. Since the traffic for maintaining the DHT is proportional to the number of finger tables to be used, the fluctuation of the traffic for maintaining the DHT can be estimated.

In addition, when f _amp is changed, how many times the query is transferred (number of recursive stages) has the same tendency if the same algorithm is used. _Therefore , f _amp is assigned to a model obtained by measurement or analysis in advance. You can ask for it. The amount of traffic consumed per inquiry depends on the recursion method, but is almost proportional to the number of recursion stages if the inquiry and response are messages of approximately the same size. Therefore, from this and the inquiry frequency, the fluctuation of the traffic of the DHT service can be estimated.

From this estimated value, the lowest cost f _amp can be calculated and selected. Further, when the quality of service is to be guaranteed, a minimum value of f _amp can be determined, and f _amp can be used within a range not below the minimum value. Furthermore, it is possible to avoid an overload of the system by setting an upper limit value and issuing an alarm to the administrator when the upper limit value is exceeded. This is because in the case of DHT, the load can be distributed by appropriately adding nodes.

Since a large change in f _amp leads to a large rewrite of the finger table, it is actually desirable to make a slight change sequentially each time fix fingers are run.

Since the find_succ process uses an old finger, the search for the vicinity (particularly the successor side) can be realized at low cost. Therefore, in particular, the cost in the direction of decreasing f _amp (in the direction of increasing the finger interval) can be significantly reduced.

Each of the above functions can be realized even if it is described as software and processed by a computer having an appropriate mechanism.
The present embodiment can also be implemented as a program for causing a computer to execute a predetermined procedure, causing a computer to function as a predetermined means, or causing a computer to realize a predetermined function. In addition, the present invention can be implemented as a computer-readable recording medium on which the program is recorded.

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. In addition, various inventions can be formed by appropriately combining a plurality of components disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

The figure which shows the structural example of the distributed hash table system which concerns on one Embodiment of this invention. The figure which shows the structural example of the DHT node which concerns on the same embodiment The figure which shows an example of the database 34 The figure which shows an example of a finger table The flowchart which shows an example of the procedure which searches the DHT node registered into a finger table The figure for demonstrating the target obtained when _famp = 1.0 The figure for demonstrating the target obtained when _famp = 2.0 The figure for demonstrating the target obtained when _famp = 3.0 Flow chart showing an example of a processing procedure when an inquiry message is received The figure which shows an example of the format of ID provided for every individual item in the individual item distribution management / traceability system

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Client, 3 ... DHT node, 8 ... Network, 31 ... Message processing part, 32 ... Own node information storage part, 33 ... Adjacent node information storage part, 34 ... Database, 35 ... Finger table, 36 ... Finger table management part 37: Communication unit, 361 ... Parameter setting unit, 362 ... Calculation unit, 363 ... Node search unit

Claims (11)

In one node device constituting a distributed hash table system composed of a plurality of node devices that share different hash value ranges,
When a message to be received by a node device that shares a specific hash value is to be transmitted without being able to identify the node device, a finger table that registers information of a plurality of node devices that are candidates for transferring the message When,
Selecting means for selecting a number of node devices to be registered in the finger table from among all the node devices other than the own node constituting the distributed hash table system, according to a designated parameter;
A different hash value is associated with each node device constituting the distributed hash table system,
The selection means is a hash value associated with the own node raised to the power of 2 (r is a real number selected from a range of 0 ≦ r <m, and m is a hash space managed by the distributed hash table system) The node device to be registered in the finger table is selected based on the hash value obtained by adding the number of bits) and the hash value associated with the node device other than its own node. A node device.   The node device according to claim 1, wherein the selection unit selects the number of node devices by changing the value of r according to the parameter.   The selection means calculates the hash value associated with its own node to the power of 2 (i · m / M) (M is an arbitrary value, and i is an arbitrary integer value in a range of 0 ≦ i ≦ M−1). And m is the number of bits in the hash space managed by the distributed hash table system.) And the hash value associated with the node device other than its own node The node device according to claim 1, wherein a node device to be registered in the finger table is selected. M is a value determined based on the parameter,
4. The node device according to claim 3, wherein the selection unit selects the node device for all i satisfying 0 ≦ i <M−1.   5. The node apparatus according to claim 3, wherein the M is given by M = f · m, where f is the parameter.   The node device according to claim 5, wherein the parameter f is determined by a request from an application.   7. The node device according to claim 1, further comprising setting means for setting the parameter.   Means for specifying one node device to which the message is to be transferred based on the specific hash value and a hash value associated with each of the plurality of node devices registered in the finger table; The node device according to claim 1, further comprising: a node device according to claim 1. The range of hash values shared by each node device constituting the distributed hash table system is a hash value of key information for data stored and managed by the node device, and the key information includes individual item distribution management and traceability It is an identifier assigned to each individual item in the system, and the data is information on the individual item to which the identifier is assigned,
9. The node device according to claim 1, wherein the message is an inquiry message designating the key information. Configure a distributed hash table system consisting of multiple node devices that share different hash value ranges, and send a message that should be received by a node device that shares a specific hash value without specifying that node device. In this case, in the transfer destination node determination method of one node device provided with a finger table in which information of a plurality of node devices that are candidates for transferring the message is registered,
A setting means provided in the node device sets a designated parameter;
The selection means included in the node device selects a node device to be registered in the finger table by the number determined according to the parameter from all the node devices other than the own node constituting the distributed hash table system. With steps,
A different hash value is associated with each node device constituting the distributed hash table system,
The selection means is a hash value associated with the own node raised to the power of 2 (r is a real number selected from a range of 0 ≦ r <m, and m is a hash space managed by the distributed hash table system) The node device to be registered in the finger table is selected based on the hash value obtained by adding the number of bits) and the hash value associated with the node device other than its own node. A forwarding node determination method. In a program for causing a computer to function as one node device constituting a distributed hash table system composed of a plurality of node devices that share different hash value ranges,
When a message to be received by a node device that shares a specific hash value is to be transmitted without being able to identify the node device, a finger table that registers information of a plurality of node devices that are candidates for transferring the message When,
The computer realizes selection means for selecting a node device to be registered in the finger table from among all the node devices other than its own node constituting the distributed hash table system according to a designated parameter. As well as for
A different hash value is associated with each node device constituting the distributed hash table system,
The selection means is a hash value associated with the own node raised to the power of 2 (r is a real number selected from a range of 0 ≦ r <m, and m is a hash space managed by the distributed hash table system) The node device to be registered in the finger table is selected based on the hash value obtained by adding the number of bits) and the hash value associated with the node device other than its own node. Program.

Images