JP5148525B2 - Content delivery method and system - Google Patents

Description

  The present invention relates to a content distribution method and system, and more particularly, to a content distribution method and system for selecting content recipients based on context information representing the attributes of the content recipient.

  At least one content distributor and peer nodes of a plurality of content receivers constitute a peer-to-peer network, each content receiver pre-advertises its context information to the peer-to-peer network, and each content distributor Non-Patent Document 1 and Patent Document 1 disclose a technique for performing a context search and distributing content only to content recipients whose context satisfies a search condition.

  The context information is one of search parameters used for the content receiver to narrow down the content desired to be distributed, and the content sender to narrow down the distribution destination of the content, and includes context items and context values. . The context item is, for example, the type of content desired by the content receiver, the installation location of the receiving node, the desired reception time of the content, the communication means used, gender, age, hobby preference, occupation type, family information, and the like. The context value is a specific value of the context item. For example, when the content type desired to be received is used as the context item, content such as “sports”, “travel”, and “music” is stored. It is a value or code indicating the type.

  On the other hand, since the context information of each content recipient distributed and managed in the peer-to-peer network does not include information on other content recipients who can expect indirect content delivery from the content recipient, the content distributor Cannot quantitatively determine and select content recipients suitable for efficient content distribution from among a plurality of content recipients that satisfy the context search conditions.

  In order to solve such a technical problem, the inventors of the present invention, for each content receiver, sets the number of other content receivers having the same context value as that of the content receiver on the peer-to-peer network. In addition to the content receiver identification information that satisfies the context search condition, the content distributor has the same context value as the content receiver and can expect indirect content distribution from the content receiver. We have also invented a technology for acquiring content related to content recipients and distributing content so that the total number of content recipients increases, and filed a patent application (Patent Document 2).

JP 2008-158931 A Japanese Patent Application No. 2008-304178

Kanno, Nakamura, "Push-type P2P Information Delivery Platform", 2007 IEICE General Conference, B-19-20, p.582, 2007.03.07

  In the technique of Patent Document 2, as shown in FIG. 8, the content distributor is transferred the number of content receivers (primary receivers) to which the content is directly distributed from the content distributor and the content receiver. Thus, the number of content receivers (secondary receivers) that receive indirectly can be predicted. However, since the content recipient (secondary recipient) who indirectly received the content may further forward the received content to another content recipient (tertiary recipient), the total number of final content recipients is managed. I couldn't. For this reason, there has been a technical problem that content is unintentionally distributed to a large number of content recipients.

  An object of the present invention is to provide a content distribution method and system capable of solving the above-described problems of the prior art and limiting the number of content recipients to a target number.

  In order to achieve the above object, according to the present invention, a content receiver pre-registers his / her context on the network, and the content distributor selectively distributes content to content receivers that match the context search conditions. In the content distribution system, a context management table for managing the context of the content recipient on the network, a means for retrieving a content recipient that matches the search condition from the context management table, and content to the content recipient that matches the search condition Means for estimating the expected number of total content recipients accumulated by transferring content distributed from the content recipient to other content recipients having the same context and repeating this, and the expected number So that approaches the target number The content distributor has a means for selecting a content receiver to which the content is distributed from content receivers that match the search conditions, and a means for distributing the content to the selected content receiver of the direct distribution destination. It is characterized by.

  According to the present invention, content is directly and indirectly received from a distributor by narrowing down the content receivers that directly distribute the content from the content receivers that are found by the context search and satisfy the desired search conditions. The total number of content recipients to be received can be brought close to the desired target number.

It is the figure which showed the structure of the peer to peer network to which this invention is applied. It is the figure which expressed typically the context table of 1st Embodiment. It is the figure which expressed typically the context management table of 1st Embodiment. It is the flowchart which showed operation | movement of 1st Embodiment. It is the figure which expressed typically the context table of 2nd Embodiment. It is the figure which expressed typically the context management table of 2nd Embodiment. It is the figure which expressed typically the context management table of 3rd Embodiment. It is the figure which showed the conventional content delivery method typically.

  Hereinafter, the best embodiment of the present invention will be described in detail with reference to the drawings. FIG. 1 is a diagram showing the configuration of a peer-to-peer network to which the content distribution method and system of the present invention is applied. At least one content distribution peer node Ps and peer nodes Pi (P0, P1, P2... ) And a plurality of context management peer nodes Pj (Pa, Pb, Pc...) That manage the context information of each content recipient xi are connected to each other in a peer-to-peer network.

  Although not shown, in this embodiment, the peer nodes of other members (content recipients) that have the same context as each content recipient xi and form an internal community etc. are also connected to the peer-to-peer network. ing. Each content recipient xi transfers the content received by the content recipient xi with a predetermined transfer probability to other content recipients whose context is known to be the same.

  In order to simplify the description, a case where each peer node specializes in one of the content distribution peer node Ps, the peer node Pi of the content recipient xi, and the context management peer node Pj will be described as an example. Each peer node may have a plurality of functions.

  Each peer node on the network has a context table 1 in which the context of each content recipient xi is registered, a context management table 2 that manages content recipients xi with the same context, and a content distribution destination based on the context. A search request unit 3 that generates a search condition for searching, describes this in a search request message, and transmits the search request message, and searches the context management table 2 for a content recipient that receives the search request message and satisfies the search condition. A search response unit 5 that returns a search response message in which a search result is recorded, a search result identification unit 6 that identifies a content recipient that satisfies the search condition, and a content recipient that matches the search condition Content delivered to the same context from the content recipient A total number estimation unit 7 for estimating the total number (expected number) N of content receivers that are transferred to other content receivers known to be transmitted with a predetermined probability and accumulated by repeating this, and based on the estimation result A direct distribution destination determination unit 8 that determines a direct distribution destination of content from among content receivers that satisfy the search condition, and a content distribution unit 9 that distributes the content to the determined direct distribution destination are distributedly arranged. Has been.

In this embodiment, as shown in FIG. 1, each context table 1 is managed in each content recipient xi's peer node Pi, and each context management peer node Pj has a context management table 2, a search unit 4, and a search response unit. 5, the search request unit 3, the search result identification unit 6, the total number estimation unit 7, the direct distribution destination determination unit 8, and the content distribution unit 9 are arranged in the content distribution peer node Ps.
[First embodiment]

  FIG. 2 is a diagram schematically showing the contents of the context table 1 according to the first embodiment of the present invention, and the context value is registered for each content receiver using the context item as an index. For example, in the context table 1 of the content recipient x0, the context value of the context item A is “a”, the context value of the context item B is “b”, as shown in FIG. An entry indicating that the context value of context item C is “c” is registered. The same applies to the context tables 1 of the other content recipients x1 and x2 shown in FIGS.

  When each content receiver x0, x1, x2 publicizes the information of its own context table 1 on the network, the context management table 2 of each context management peer node Pa, Pb, Pc has a context value and a context value for each context item. Identification information (recipient ID) of each content receiver is registered. FIG. 3 is a diagram showing an example of the context management table 2, in which a recipient ID is registered using a context value as an index.

  Next, the operation of the present embodiment will be described with reference to a flowchart. FIG. 4 is a flowchart showing the operation of the content distribution peer node Ps in this embodiment.

  In step S1, the search request unit 3 of the content distribution peer node Ps sets a search condition for specifying a content distribution destination. Here, the description will be continued assuming that “the context value of context item A is a” is set as a search condition. In step S2, a search request message in which the search conditions are described is transmitted over the network.

  The search unit 4 of each context management peer node Pj searches each context management table 2 based on the search condition described in the search request message. If the entry satisfying the search condition is already registered, a search response message in which the recipient ID of the entry is described as a search result is generated and returned by the search response unit 5. Here, the receiver IDs of the three members x0, x1, and 02 registered with the context value “a” as an index by the search unit 4 that searches the context management table 2 that manages the context of the context item A. A search response message in which is described is returned.

  In the content distribution peer node Ps, when the search response message is received in step S3, the process proceeds to step S4. In the search result identification unit 6, three content recipients x0, x1, x2 satisfying the search condition are identified. The In step S5, the content directly distributed to the content recipients that match the search conditions in the total distribution estimation unit 7 is further transmitted from the content recipients to other content recipients whose context is the same. The expected number N of the total number of content recipients who directly and indirectly receive the content, which is transferred and indirectly received and accumulated by repeating this, is calculated.

  Here, the number n of content recipients to which content is directly delivered from the content delivery peer node Ps, and each content recipient forwards the received content to other content recipients whose context is known to be identical to itself Given the probability f, the probability p that each peer node is active, and the average number d of content recipients that each content receiver knows and has the same context as the content recipient, the total number of content recipients The expected number N is given by the following equation (1) as the sum of infinite geometric series.

  Therefore, for example, when the content transfer probability f = 0.5, the peer node activation probability p = 0.25, and the average number of content recipients d = 4.0, the number of content recipients n that directly receive content is one. In this case, the expected number N of content distribution is two people. Similarly, if n = 2, the expected number N = 4, and if n = 3, the expected number N = 6. The content transfer probability f, the peer node activation probability p, and the content recipient average number d are separately determined based on past empirical rules, and are given in advance to the content distribution peer node Ps.

In step S6, an appropriate number of content recipients are directly selected as distribution destinations from among a plurality of content recipients that satisfy the context search condition so that the expected number N of content recipients approaches the target number Nt. For example, if it is desired to distribute content to four content recipients having the context value a, two arbitrary recipients are selected from the three content recipients x0, x1, and x2 that satisfy the context search condition. . In step S7, the content is directly distributed to the two selected content recipients.
[Second Embodiment]

  Next, a second embodiment of the present invention will be described. FIG. 5 is a diagram schematically representing the registration contents of the context table 1 in the present embodiment, in which context values are registered for each content receiver using the context item as an index, and the context value is the same for the context item. The number of other content recipients known to exist (the number of known recipients) is registered.

  For example, in the context table 1 of the content recipient x0 shown in FIG. 5A, the context value of the context item A is “a”, and the context item A has the same context value “a” as itself. An entry indicating that the number of known content recipients “10” is “10” is registered. Similarly, for the context item B, an entry is registered indicating that the number of known content recipients is “2”, the context value is “b”, and the context value is “b”. . Similarly, for the context item C, an entry indicating that the number of known content recipients having the context value “c” and the context value “c” is “1” is registered. . The same applies to the context table 1 of the content recipient x1 shown in FIG. 5B and the context table 1 of the content recipient x2 shown in FIG.

  When each content receiver x0, x1, x2 advertises the information of its own context table 1 on the network, the context management table 2 of each context management peer node Pa, Pb, Pc has a context for each context item. A value, identification information (recipient ID) of each content receiver, and the number of known receivers are registered.

  FIG. 6 is a diagram showing an example of the context management table 2 in the present embodiment, in which the context ID is used as an index and the number of known recipients whose context value is the same as the content receiver is associated with the context ID. ing.

  In the context management table 2 that manages the context value of the context item A, the number of known recipients whose context value is “a” is “10” for the content recipient x0, as shown in FIG. An entry indicating that “5 people” are registered for the content recipient x1 and “7 people” is registered for the content recipient x2.

  Similarly, in the context management table 2 that manages the context value of the context item B, the number of known recipients with the context value “b” is “2” for the content recipient x0, as shown in FIG. An entry to the effect that “5 people” is registered for “person” and content recipient x1 is registered. Similarly, in the context management table 2 that manages the context value of the context item C, the number of known recipients whose context value is “c” is “1” for the content recipient x0, as shown in FIG. An entry indicating that “person”, “3” for the content recipient x1, and “2” for the content recipient x2 is registered.

  In this configuration, in the present embodiment, the search request unit 3 of the content distribution peer node Ps describes “the context value of the context item A is a” as a search condition, as in the first embodiment. When the search request message is transmitted to the network, the search response in which the receiver IDs of the three members x0, x1, 02 registered with the context value “a” as an index and the number of known receivers are described. A message is returned from the context management peer node Pa.

  In the distribution total number estimation unit 7 of the content distribution peer node Ps, for all combinations of the content receivers that satisfy the search condition, the content distributed to all the content receivers included in the combination is known from the content receivers. The expected number N of the total number of content recipients who directly and indirectly receive the content, which is transferred to the user and accumulated by repeating this, is calculated.

  Here, the number m of content recipients included in the combination of content recipients, the total number n of known recipients of each content recipient included in the combination of content recipients, the content transfer probability f of each peer node, and each peer node Given the activation probability p and the average number d of content recipients that are known to each content recipient and have the same context as that of the content recipient, the total number of content recipients is not considered if duplicate content distribution is not considered The expected number N is calculated by the following equation (2).

  As shown in FIG. 6, the number of other content receivers (number of known recipients) having the context value a known by the content receivers x0, x1, and x2 is 10, 5, and 7, respectively. If so, m = 2 and n = 15 for the combination of the content recipients x0 and x1. Similarly, in the combination of the content recipients x1 and x2, m = 2 and n = 12. Similarly, in the combination of content recipients x2 and x0, m = 2 and n = 17. Therefore, for example, if f = 0.5, p = 0.25, d = 4.0 is given, if the content is directly delivered to the combination of the content recipients x0 and x1, the expected number N of the total number of content recipients Will be 5.75.

  Similarly, if the content is directly distributed to the combination of the content receivers x1 and x2, the expected number N is 5 and if the content is directly distributed to the combination of the content receivers x2 and x0, the expected number N is 6.25. Become.

As described above, when the calculation of the expected number N is completed for all the combinations of content receivers that satisfy the context condition, the combination of content receivers whose expected number N is closest to the target value Nt of the content receiver It is determined by the direct delivery destination determination unit 8. The content distribution unit 9 directly distributes content to the combination of content recipients. For example, when content is to be distributed to five content recipients having the context value a, the content owner selects the content recipients x1 and x2 from the content recipients obtained by the context search and directly transmits the content. To deliver.
[Third embodiment]

  Next, a third embodiment of the present invention will be described. In the first and second embodiments described above, duplication of content receivers that receive content not indirectly from the content distributor but indirectly from other content receivers is not considered. Therefore, in the third embodiment of the present invention, the number of content recipients to which content is directly distributed is determined in consideration of overlapping content distribution.

  In this embodiment, each content recipient xi that makes up a peer-to-peer network advertises to the peer-to-peer network, along with its context information, the identifiers of other content recipients whose context is known to be the same. The information related to the content receiver ID advertised is distributed and managed on the peer-to-peer network together with its content receiver ID, for example, using the context value as an index.

  FIG. 7 is a diagram showing an example of the context management table 2. Here, there are three content recipients x0, x1, and x2 having the context value c in the context item C, and the content recipient x0 knows. The other content receiver with context value c is one of x5, and the other content receiver with context value c is known by three, x2, x3, and x4. There are two other content recipients x1 and x3 having a context value c that the content recipient x2 knows.

  In this embodiment, a search request message in which “the context value of context item C is c” is described as a search condition from the search request unit 3 of the content distribution peer node Ps, as in the first and second embodiments. Is transmitted over the network, a search response message in which the receiver IDs of the three members x0, x1, 02 registered with the context value “c” as an index and the IDs of the known receivers are described. Returned from the context management peer node Pc.

  In the distribution total number estimation unit 7 of the content distribution peer node Ps, for all combinations of the content receivers that satisfy the search condition, the content distributed to all the content receivers included in the combination is known from the content receivers. The expected number N of all the content recipients who directly and indirectly receive the content, which is transferred to the user and accumulated by repeating this, is calculated.

  Here, the number m of content recipients included in the combination of content recipients, and content reception excluding duplicate content recipients among other content recipients known to the content recipients included in the combination of content recipients The same, given the number n of users, the content transfer probability f of each peer node, the activation probability p of each peer node, and the average number d of content receivers that each content receiver knows and has the same context The expected number N of the total number of content recipients taking into account the duplicate reception of content by the content recipients is calculated by the following equation (3).

  However, symbol Bi is the probability that a content recipient who indirectly receives content by i transfers will receive the content redundantly, and is given an expected value M of the total number of content recipients having the same context value. For example, the recurrence formula of the following formula (4) is given.

  For example, in FIG. 7, m = 2 in the combination of content recipients x0 and x1, and duplicate content recipients are excluded from other content recipients having the context value c known by the content recipients x0 and x1. Since there are four content recipients x2, x3, x4, and x5, n = 4.

  Similarly, in the combination of the content receivers x1 and x2, m = 2, and other content receivers having the context value c known by the content receivers x1 and x2, the content receivers excluding duplicate content receivers Since n is x3 and x4, n = 2. Similarly, the content recipients excluding duplicate content recipients among other content recipients having the context value c known by the content recipients x2 and x0 and m = 2 and the content recipients x2 and x0. Since x1, x3, and x5, n = 3.

  Therefore, if f = 0.5, p = 0.25, d = 4.0, and M = 1000, the expected number of all content receivers that receive the content when the content is delivered directly to the combination of content recipients x0 and x1 N is about 3.0. Similarly, the expected number N when the content is directly distributed to the combination of the content receivers x1 and x2 is about 2.5 people. Similarly, the expected number N when the content is directly distributed to the combination of the content receivers x2 and x0 is about 2.75.

  The direct distribution destination determination unit 8 selects an appropriate combination of content recipients that satisfy the context search condition so that the expected number N of the total number of content recipients approaches the desired target number Nt. The content distribution unit 9 directly distributes the content to the content receivers of the selected combination. For example, if the target number Nt of the total number of content recipients is 3, the direct delivery destination determination unit 8 directly delivers the content to the combination of content recipients x0 and x1.

  By the above procedure, the content distributor receives content indirectly by considering the probability of receiving duplicate content by appropriately selecting the combination of content recipients that directly distribute the content. The expected number N of the total number of content recipients including the user can be brought close to the desired total number of content recipients Nt.

  If the number of content transfers is limited, the content owner will set the expected number N of all content recipients including indirect content recipients due to content transfer when there are restrictions on the number of content transfers. calculate.

  For example, when the maximum allowable content transfer count I is given, instead of the calculation formulas (1), (2), (3) of the expected number N of content receivers in each embodiment, the following formulas (5), (6) and (7) may be used. In this way, the content owner can bring the expected number of content recipients closer to the desired target number Nt while limiting the number of times the content is transferred.

DESCRIPTION OF SYMBOLS 1 ... Context table, 2 ... Context management table, 3 ... Search request part, 4 ... Search part, 5 ... Search response part, 6 ... Search result identification part, 7 ... Total number estimation part, 8 ... Direct delivery destination determination part, 9 ... Content distribution department

Claims (7)

In a content distribution system in which a content receiver pre-registers his / her context on the network, and the content distributor selectively distributes content to a content receiver that matches the context search conditions.
A context management table for managing the context of content recipients on the network;
Means for searching for content recipients matching the search condition from the context management table;
Content distributed from content distributors to content receivers that match the search conditions is transferred from the content receivers to other content receivers with the same context, and this is repeated until the total number of content receivers accumulated A means of estimating the expected number;
Means for selecting a content receiver from which a content distributor directly distributes content from among content receivers that match a search condition so that the expected number approaches the target number;
A content distribution system comprising: means for distributing content to a content recipient of the selected direct distribution destination. Means for obtaining a transfer probability that the received content is transferred from the content receiver to another content receiver having the same context;
The content distribution system according to claim 1, wherein the estimation unit estimates the expected number based on the transfer probability. Means for obtaining a transfer probability that the received content is transferred from the content receiver to another content receiver having the same context;
In the context management table, the content recipient is associated with the number of other content recipients having the same context as the content recipient.
The search means searches for the number of content receivers that match the search condition and the number of other content receivers having the same context as the content receivers,
The content distribution system according to claim 1, wherein the estimation unit estimates the expected number based on the number of other content recipients having the same transfer probability and the same context. Means for obtaining a transfer probability that the received content is transferred from the content receiver to another content receiver having the same context;
In the context management table, the content receiver is associated with identification information of another content receiver having the same context as the content receiver,
The search means searches for identification information of content receivers that match the search conditions and other content receivers having the same context as the content receivers,
2. The content distribution according to claim 1, wherein the estimation unit estimates a non-overlapping expected number of content receivers based on identification information of other content receivers having the same transfer probability and the same context. system. Means for limiting the number of times a content recipient forwards received content to another content recipient;
5. The content distribution system according to claim 2, wherein the means for estimating the expected number estimates the expected number based on the limited number of transfers.   The transfer probability is known by each content receiver, the probability f that the received content is transferred to another content receiver whose context is the same as that of itself, the probability p that each peer node is activated, and each content receiver knows 6. The content distribution system according to claim 2, further comprising at least one of an average number d of content recipients having the same context as the content receiver. In a content distribution method in which a content receiver pre-registers his / her context on the network and the content distributor selectively distributes content to a content receiver that matches the context search condition.
Expected number of total number of content recipients accumulated by the content delivered directly to the content recipients that match the search criteria being transferred from the content recipient to other content recipients with the same context and repeated A procedure for estimating
A procedure for selecting a direct delivery destination of content from among content recipients that match the search condition so that the expected number approaches the target number;
And a procedure for delivering content to the selected content recipient at the direct delivery destination.

Images