JPH1185604A - Contents arrangement changing method and system and storage medium storing contents arrangement change program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method and a system for changing contents arrangement and a storage medium storing contents arrangement change program with which the arrangement of contents can be dynamically changed before exceeding a critical value processible for a media server by predicting the load change of the media server to fluctuate in real time in accordance with the change of access from contents display terminal equipment. SOLUTION: Concerning contents stored at the media server, location information, sharpness information and information of access from terminal equipment are collected (S1), and when requests are generated from a lot of terminal equipment to the contents of the specified media server, the contents as the factor of access load are specified (S2). Then, the condition of access from the terminal equipment to the specified contents is analyzed (S3) and time exceeding the critical value of access load processible for the media server is predicted (S4). Before the access load of the media server reaches the critical value, the copy of contents is prepared or the location is rearranged by the relevant and other contents (S5).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method and system for changing the arrangement of contents and a storage medium storing a program for changing the arrangement of contents, and more particularly to providing continuous media contents represented by video and audio in real time each time a request is made. The present invention relates to a content arrangement change method and system in a multimedia service system for performing a content arrangement change such that a load is not biased to a specific media server, and a storage medium storing a content arrangement change program.

[0002]

2. Description of the Related Art In a technology realized by a conventional content arrangement changing apparatus, an access frequency and a call blocking rate obtained from a result of collecting statistical information are used as a trigger for executing an arrangement change. In addition, access frequency information is categorized by genre, price,
There is a method of estimating an access occurrence interval from the popularity and using it for content layout design.

[0003]

However, in the method of changing the content arrangement using the access frequency and the call blocking rate obtained from the result of collecting the statistical information, an appropriate content arrangement cannot be obtained until a past history is obtained. Therefore, it is difficult to immediately respond to a change in the load of the media server that changes in real time.

Further, a method of estimating an access occurrence interval from a genre, a price and a popularity based on the access frequency and using the estimated access interval for the content layout design is as follows. There is a problem in that it is not credible to determine whether to make a copy or rearrange the location as a layout change for a content having a long time.

[0005] The present invention has been made in view of the above points, and predicts a load change of a media server which fluctuates in real time in accordance with a change in access from a content display terminal device, and sets a limit value at which the media server can process. Before and after, it is necessary to dynamically change the content arrangement. In particular, a content arrangement change method and system capable of reflecting an access pattern that differs depending on the service type in the prediction, and a storage medium storing the content arrangement change program. The purpose is to provide.

[0006]

FIG. 1 is a diagram for explaining the principle of the present invention. The content arrangement change method of the present invention is a method of performing a content arrangement change such that a load is not biased to a specific media server in a media server that provides continuous media content represented by video and audio in real time each time a request is made. Yes, if multiple terminal devices request content on the same media server, copy the content to a different media server or relocate the location to distribute the load on the media server At this time, location information, freshness information, and access information from a terminal device are collected for the content stored by each media server connected to the local communication network or the wide area communication network (step 1), and a specific media server is collected. Access request when a large number of terminal devices request The content causing the cause is specified (step 2), the access status of the specified content from the terminal device is analyzed (step 3), and the time exceeding the limit value of the access load that can be processed by the media server is predicted (step 2). 4) Before the access load of the media server reaches the limit value, a copy of the content is created or the location is relocated between the content and other content (step 5).

According to the present invention, when collecting access information (step 1), log information is received from the media server group, and the access occurrence time, the service continuation time and the access request source terminal device are obtained from the log information. Extract information about whereabouts. Further, the present invention uses a predetermined criterion when specifying the content causing the access load (step 2), and determines that the access load is caused when the content exceeds the criterion.

Further, according to the present invention, when predicting the time exceeding the access load limit value (step 4), an access frequency model and a service duration model are constructed for each content and each service type from the past access history. Then, the access frequency model, the service duration model, and the log information are evaluated. Further, the access frequency model for each content described above totals the access occurrence intervals based on the access occurrence time, and the access continuation model for each service type averages the access occurrence intervals of all contents belonging to the service type. The service duration model for each content uses the distribution of how long the service has been displayed at a time rate with respect to the total display time of the content. The service duration model for each service type is For all contents belonging to the type, values obtained by multiplying the service duration model by the number of accesses are added, and the distribution obtained by dividing by the total number of accesses is used.

When estimating the time exceeding the limit value of the access load (step 4), for the currently occurring access, the access time for continuing the service is estimated from the service continuation time model, and the access time is newly calculated. Regarding the access that occurs, the access time is estimated from the service duration model, assuming that the access occurs according to the access frequency model, and the load is predicted by integrating them.

[0010] The present invention also relates to a novel content or
For content whose access number is not sufficient, a weight is assigned to the access frequency model based on freshness information included in log information of the content. In addition, the present invention evaluates log information, an access frequency model, and a service duration model, and issues a content arrangement change to a media server to be operated by a media server group based on the evaluation.

FIG. 2 is a diagram showing the principle of the present invention. The content arrangement change system according to the present invention includes a plurality of media servers 200 that deliver contents accompanied by continuous media represented by video and audio upon request, and a plurality of media servers 200 that access the media servers 200 and display the contents. A multimedia service system comprising a terminal device 300 and a local communication network for transferring content between the media server 200 and the terminal device 300 in real time, or
A system in which a plurality of multimedia service systems are connected via a wide area communication network that transfers contents in real time, wherein a content stored in a local communication network or a media server 200 connected to the wide area communication network is stored. An access information collection unit 110 for collecting location information, freshness information, and access information from terminal devices; and a content that causes an access load when a large number of terminal devices request a content of a specific media server 200. Load cause content specifying means 120 to specify
Access information analyzing means 130 for analyzing access information from the terminal device to the content specified by the load cause content specifying means 120;
A limit value predicting unit 140 for predicting a time when the access load exceeds a limit value of the access load which can be processed by the media server 200;
Before the access load of the content reaches the limit value, the content arrangement changing apparatus 100 includes a copy / relocation unit 150 for creating a copy of the content or relocating the location between the content and another content.

The above-mentioned access information collecting means 110
Receives log information from a group of media servers and extracts from the log information information regarding the access occurrence time, the service continuation time, and the location of the terminal device requesting access. The load cause content specifying means 120 uses a predetermined criterion, and determines that the access load is a cause when the criterion is exceeded.

The limit value predicting means 140 constructs an access frequency model and a service duration model for each content and each service type from the past access history, and evaluates the access frequency model, the service duration model, and the log information. Evaluation means. In addition, the access frequency model for each of the contents tabulates the access occurrence intervals based on the access occurrence time, and the access continuation model for each service type calculates the average of the access occurrence intervals of all the contents belonging to the service type. The service duration model for each content uses the distribution of how long the service has been displayed at a percentage of the total display time of the content. The service duration model for each service type is For all contents belonging to the type, values obtained by multiplying the service duration model by the number of accesses are added, and the distribution obtained by dividing by the total number of accesses is used.

The above-mentioned limit value predicting means 140 estimates the access time for continuing the service from the service continuation time model for the currently occurring access, and calculates the access frequency model for the newly occurring access. Means for estimating the access time from the service duration model assuming the occurrence of access according to the above, and predicting the load by integrating them.

[0015] The present invention also relates to a novel content or
For content whose access count is not sufficient, a means for weighting the access frequency model based on freshness information included in the log information of the content is included. The above-mentioned duplication / relocation means 150 evaluates the log information, the access frequency model, and the service duration model, and issues a content allocation change to the media server to be operated by the media server group based on the evaluation. Including means.

The storage medium storing the content arrangement change program according to the present invention can be used in a plurality of media servers which transfer real-time contents including continuous media such as video and audio in response to a request from a terminal device. The content arrangement is changed so that the load is not biased to the media server of the present invention, and the location information, freshness information, and access information from the terminal device of the content stored in each media server are transmitted to the local communication network or the wide area communication. An access information collection process for collecting contents via a network, a load cause content specifying process for specifying contents that cause an access load when a large number of terminals request contents of a specific media server, and a load cause Terminal device for the content specified by the content specifying process An access information analysis process for analyzing the access information of the media server, a limit value prediction process for predicting a time exceeding a limit value of an access load that can be processed by the media server, and a corresponding content before the access load of the media server reaches the limit value. And a copy / relocation process for relocating the location between the content and other content.

The access information collecting process receives log information from the media server group, and extracts information on the access occurrence time, the service continuation time, and the location of the access requesting terminal device from the log information. In addition, the load cause content identification process described above includes:
Using a predetermined criterion, if the criterion is exceeded, the access load is caused.

The above-described limit value prediction process estimates the access time for continuing the service from the service continuation time model for the currently occurring access and the access frequency model for the newly occurring access. Assuming the occurrence of the access, a process of estimating the access time from the service duration model and integrating them to predict the load is included.

In addition, a process is provided for assigning weights to the access frequency model based on the freshness information included in the log information of the new content or the content of which the number of accesses is insufficient. In addition, the above-mentioned duplication / relocation process evaluates log information, an access frequency model, and a service duration model, and issues a content arrangement change to a media server to be operated by a media server group based on the evaluation. Process.

As described above, according to the present invention, not only an access frequency model but also a service duration model is constructed for each content and each service type from the past access history, and the media server access that matches the actual situation as much as possible is achieved. By predicting the load and initiating the content placement change process before it reaches the limit, call loss and performance degradation due to access beyond the performance of the media server is reduced.

Further, according to the present invention, for the content stored in each media server, information on the access occurrence time, the service continuation time, and the location of the access requesting terminal device are collected. For new content or content with insufficient access count, a model for each service type is used to ensure the reliability of the access frequency model and the service duration model. In particular, for new content, a high access frequency is often expected, so the access frequency model is weighted based on the freshness information of the content.

The access frequency model is constructed from log information relating to the access occurrence time for each content. The access frequency is collected as information for each time zone, as in Japanese Patent Application No. 8-307853. The above-mentioned access frequency model for each content tabulates the access occurrence intervals based on the access occurrence time, and the access continuation model for each service type takes the average of the access occurrence intervals of all the contents belonging to the service type. The service duration model for each content uses the distribution of how long the service has been displayed at a time rate with respect to the total display time of the content. The service duration model for each service type is For all contents belonging to the type, the value obtained by multiplying the service duration model by the number of accesses is added, and the distribution obtained by dividing by the total number of accesses is used.

Further, in the present invention, the load prediction of the media server predicts the time exceeding the limit value of the access load. For the access that is currently occurring, the service is continued from the service duration model. The access time is estimated, and for newly occurring accesses, the access time is estimated from the service duration model, assuming that the access occurs according to the access frequency model, and these are integrated to obtain the time of the media server load evaluation value. Calculate the change.

Further, according to the present invention, the content arrangement change process is started at a timing when it is predicted that the access load of a certain media server exceeds the limit. When the content arrangement change process is started, a copy of the content is created and the location of the content is rearranged. The content of the operation is determined by using an activation process selection algorithm.

As described above, in the present invention, log information of the media server is analyzed, and an access model is constructed for each content and each service type. Using this model, a change in the access load is predicted for each media server, and before the media server exceeds its limit value, the duplication of the content causing the load and the relocation of the content location are performed.

[0026]

FIG. 3 shows the configuration of a service system to which a content arrangement changing device according to the present invention is connected.
The system shown in FIG. 1 includes a plurality of media servers 1A-1,... 1A-n for delivering contents in response to a request, a contents arrangement change device 2A, 2B, 2C, and a plurality of terminal devices 4A-1 for displaying contents. ~ 4A-n, 4B-1
4B-n, 4C-1 to 4C-n, and local communication networks 3A, 3B, 3C for transferring contents between the media server group 1 and the terminal device group 4 in real time. For example, a content arrangement change device 2A for the media server group 1A is connected via a local communication network 3A. When a plurality of systems A, B, and C are connected by the wide area communication network 10, as shown in the figure, the content arrangement change apparatuses 2A, 2B, and 2C are connected to the respective local communication networks 3A, 3B, and 3C.
It may be connected to 3C or to a wide area communication network.

FIG. 4 shows the configuration of the content arrangement changing device of the present invention. Content arrangement changing device 2 shown in FIG.
Is composed of a main control unit 20 and a database 25. The main control unit 20 includes a log information analysis unit 21, a modeling unit 22,
It comprises a load prediction unit 23 and a startup process selection unit 24. The main control unit 20 of the content arrangement change device 2 receives the log information from the media server group 1, extracts necessary information by the log information analysis unit 21, and stores it in the database 25. From this information, the modeling unit 22 constructs an access frequency model and a service duration model, stores them in the database 25, constructs log information and a model, and stores them in the database 25. It evaluates using the model, and issues a content arrangement change instruction guided by the algorithm of the activation process selection unit 24 to the media server that is the operation target of the media server group 1.

[0028]

Embodiments of the present invention will be described below with reference to the drawings. The following embodiment will be described with reference to FIGS. First, when requests from the terminal devices 4A, 4B, and 4C in FIG. 3 are concentrated on the media server 1A-1, the content relocation device 2A relocates the content to the media server 1A-2.

First, the main control unit 20A of the content arrangement changing device 2A receives log information from the media servers 1A-1, 1A-2,..., 1A-n of the media server group 1A, and analyzes the log information. Transfer to 21A. The log analysis unit 21A extracts information necessary for analysis from the log information and stores the information in the database 25A. The modeling unit 22A configures an access frequency model and a service duration model based on the information stored in the database 25A and stores the model in the database 25A.

The load prediction unit 23A uses the log information and the model constituted by the modeling unit 22A to execute the media server 1
The load on A is evaluated and notified to the activation process selection unit 24A. The activation process selection unit 24A issues a content arrangement change instruction guided by an algorithm to the media server 1A-2 to be operated.

An example in which log information is analyzed by the content arrangement changing device 2 will be described. FIG. 5 shows an example of analyzing log information according to an embodiment of the present invention. From the log information received from the media server 1 in the same service system, using the content name as a key, the storage destination media server name, service type, freshness information, access requesting terminal device, access occurrence time, service duration, and Obtain information on service provision status.

FIG. 6 shows an example of an access occurrence time distribution and an example of an access request distribution according to an embodiment of the present invention. The figure shows an access occurrence model obtained from the access occurrence time and the information of the access request source. The time period is divided, for example, in one-hour units, the access frequency in each time period is measured, and the access request distribution, which is the ratio of access for each communication network to which the terminal device of the access source belongs, is calculated in advance.

An access generation model for each service type is as follows:
The sum of the access occurrence models for all contents belonging to the service type is calculated, and the result is divided by the number of contents (average of access occurrence intervals). The service duration distribution for each content is a distribution indicating how long the service has been displayed at a time ratio with respect to the display time of the entire content. The service duration model for each service type is the sum of the service duration model and the number of accesses weighted for all contents belonging to the service type.

FIG. 7 shows an example of two types of service duration distributions according to an embodiment of the present invention. This figure shows a service duration model for two contents having different service types. The two models shown in the figure are charged for each title, and a video-on-demand service (FIG. 7 (A)) in which the cancellation time is determined, and a free video shopping service comprising a collection of several types of topics. (FIG. 7B).

The load prediction unit 23 predicts an access load for each media server. In the media server 1, a threshold access load for starting the content arrangement change process is determined in advance for the limit value of the access load. The load prediction unit 23 extracts information on the content stored in the target media server from the database 25 of the content arrangement change device 2. Prediction is performed based on information on the total number of accesses, the access occurrence model, the service duration model, the information freshness, and the service type, in addition to the number of real-time accesses in content units. The access occurrence rate is obtained from the access occurrence model and the access continuation rate is obtained from the service continuation time model, and the change in the number of accesses is predicted. FIG. 8 is an example of content information used for load estimation according to an embodiment of the present invention.

The access load is given by a function with respect to time. FIG. 9 is an example of load prediction according to one embodiment of the present invention. For contents for which a sufficient number of accesses has been observed, the access occurrence rate and access continuation rate for each content are used. In the case of new content or content with a small number of accesses, the access load is estimated based on a service duration model, assuming that access occurs according to an access frequency model, and the load is predicted by integrating them.
That is, it is calculated by adding the content freshness information to the access occurrence rate and access continuation rate for each service type.

The method of calculating the access load will be described below. -Measure the number of current accesses and the elapsed time from the start of each access:-Calculate the service duration of the current access from the access continuation rate:-Estimate new accesses from the access occurrence rate (with fluctuations in the average access occurrence time) Calculates the service continuation time of new access from the access continuation rate: Determines the sum of the calculated accesses and determines the access load function for each content: From the sum of the functions, determine the access load function for each media server: Monitor whether the access load function of the media server exceeds a threshold value: Next, the activation process selection unit 24 activates the content arrangement change process. An example of the algorithm at the time will be described.

FIG. 10 is a flowchart for selecting an activation process for changing the content arrangement according to one embodiment of the present invention. Step 101) The load prediction unit 23 acquires a load prediction result. Step 102) It is determined whether or not the load prediction result has exceeded a predetermined threshold. If it has, the process proceeds to step 103, and if not, the process proceeds to step 101.

Step 103) The load of the content stored in the server is evaluated. Step 104) If the cause content is specified, the process proceeds to step 105; otherwise, the process proceeds to step 109. Step 105) The following judgment is made for each of the cause contents.

Step 106) It is determined whether or not there is a bias in the requesting terminal device. Step 1 if there is a bias
08, and if not, the process proceeds to step 107. Step 107) Issue a copy command in the current communication network. Step 108) Issue a copy command to the communication network where the requests are biased. Step 109) The following determination is made for the entire contents of the server.

Step 110) It is determined whether or not there is a bias in the requesting terminal device. If so, the process proceeds to step 111; otherwise, the process proceeds to step 112. Step 111) A relocation instruction is issued to the communication network in which requests are biased. Step 112) Issue a relocation instruction in the current communication network. When the activation process selecting unit 24 executes the above-mentioned content arrangement change command, for example, the content causing one or a plurality of loads may be determined from the application of each content to the predicted value of the load of the media server likely to exceed the limit value. ,
The content is specified based on criteria such as “content that contributes to load exceeding 40%”. In addition, the access bias for each local communication network is checked from the distribution of access requests to the corresponding media server and its contents. As a result, the executed step 1
07, step 108, and the contents of step 111 and step 112 are as follows.

For the content that is identified as the cause of the load and has a biased access, a copy of the content is created on the media server connected to the local communication network with the biased access and having the lowest load evaluation value (step 108). ). For the content that is specified as the load cause and the access is not biased, a copy of the content is created on the media server having the lowest load evaluation value in the local communication network to which the corresponding media server is connected (step 107).

When the cause of the load is not specified and the access is uneven, the location of the content is re-located between the media server connected to the local communication network having the access unevenness and having the lowest load evaluation value and the corresponding media server. The arrangement is executed (step 111). -If the cause of the load is not specified and the access is not biased, the location of the content is relocated between the media server with the lowest load evaluation value and the relevant media server in the local communication network to which the relevant media server is connected. Is executed (step 112).

When there are a plurality of contents to be copied in the same local communication network, the contents are copied as shown in FIG. FIG. 11 is a flowchart of the content copy processing according to one embodiment of the present invention. Step 201) The following processing is performed in descending order of the content load. Step 202) The server with the lowest load is set as the copy destination media server.

Step 203) It is determined whether or not the capacity of the copy destination media server is sufficient. If the capacity is sufficient, the process proceeds to step 205. If not, the server is excluded from the candidates in step 204. To 202. Step 205) Copy the contents of the copy source media server. If the cause of the load has not been identified, the content is rearranged based on the flow of FIG.

FIG. 12 is a flowchart of the process of rearranging the content location according to one embodiment of the present invention. Step 301) The media server with the lowest load is selected as a server to be relocated. Step 302) Select a content group candidate to be exchanged.

Step 303) It is determined whether the capacity of the server to be relocated is sufficient. If the capacity is sufficient, the process proceeds to step 304;
Move to Step 304) The content is relocated between the relocation target server and the current server.

As described above, in the content location relocation processing, a set of content that can be exchanged in terms of capacity so that the load becomes as equal as possible is obtained and executed. This allows
Estimate the access occurrence interval model and service duration model for each service type from past access pattern cases, make reliable predictions of changes in the load on the media server, and the state before the media server reaches the limit that can be processed By realizing the real-time access load fluctuation, the content arrangement change process is started.

The above embodiment of the present invention is constructed as a program (software), and is connected to a computer used as a content arrangement changing device in each local communication network, a floppy disk, a CD-ROM, or the like.
The program may be stored in a portable storage medium such as a ROM, and may be installed and executed as needed. It should be noted that the present invention is not limited to the above-described embodiment, but can be variously modified and applied within the scope of the claims.

[0050]

As described above, according to the present invention, by predicting a change in load by modeling a service duration, contents having a high access frequency but a short service time per access, and an access frequency being low. Enables a more accurate assessment of the load on content that has a long service time per access, or content that follows a unique pattern of service time per access.

In addition, by predicting the load using an access frequency model and a service duration model corresponding to the service type in consideration of the freshness information of the content, a more accurate evaluation of the load relating to the content with a small number of accesses can be performed. It becomes possible. In addition, before the access load exceeds the media server's limit value, duplication of content and rearrangement of locations that cause the load may reduce the risk of the media server deteriorating the service quality or causing a call loss. Can be reduced.

[Brief description of the drawings]

FIG. 1 is a diagram for explaining the principle of the present invention.

FIG. 2 is a principle configuration diagram of the present invention.

FIG. 3 is a configuration diagram of a service system to which the content arrangement change device of the present invention is connected.

FIG. 4 is a configuration diagram of a content arrangement change device of the present invention.

FIG. 5 is an example of analyzing log information according to an embodiment of the present invention.

FIG. 6 shows an example of an access occurrence time distribution and an example of an access request distribution according to an embodiment of the present invention.

FIG. 7 is an example of two types of service duration distributions according to an embodiment of the present invention.

FIG. 8 shows content information used for predicting a load according to an embodiment of the present invention.

FIG. 9 is an example of load prediction according to one embodiment of the present invention.

FIG. 10 is a flowchart illustrating a process of selecting a content arrangement change according to an embodiment of the present invention.

FIG. 11 is a flowchart of a content copy process according to an embodiment of the present invention.

FIG. 12 is a flowchart of relocation of a content location according to an embodiment of the present invention.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Media server, media server group 2 Content allocation change device 3 Local communication network 4, 5, 6 terminal device 20 Main control unit 21 Log information analysis unit 22 Modeling unit 23 Load prediction unit 24 Start-up process selection unit 25 Database 100 Content allocation Change device 110 Access information collecting means 120 Load cause content specifying means 130 Access information analyzing means 140 Limit value predicting means 150 Duplicating / relocating means 200 Media server group 300 Terminal device group

────────────────────────────────────────────────── ───

[Procedure amendment]

[Submission date] September 19, 1997

[Procedure amendment 1]

[Document name to be amended] Drawing

[Correction target item name] Fig. 6

[Correction method] Change

[Correction contents]

FIG. 6

Claims (22)

[Claims] 1. A multimedia service system for providing a continuous media content represented by video and audio in real time each time a request is made, and performing a content arrangement change such that a load is not biased to a specific media server. In the content arrangement change method, in the case where a request is issued from a plurality of terminal devices to the content of the same media server, the content is duplicated or located on a different media server for the purpose of distributing the load generated on the media server. When performing the relocation of the local communication network or the content stored in each media server connected to the wide area communication network,
Collects location information, freshness information and access information from terminal devices, identifies content that causes access load when a large number of terminal devices request content on a specific media server, and identifies the identified content Analyzing the access status from the terminal device to the server, predicting the time exceeding the limit value of the access load that the media server can process, and before the access load of the media server reaches the limit value, A content arrangement change method, which comprises creating or relocating a location between the content and another content. 2. The method according to claim 1, wherein when collecting the access information, log information is received from a group of media servers, and information relating to an access occurrence time, a service continuation time, and a location of an access requesting terminal device is extracted from the log information. Item 1. The content arrangement change method according to Item 1. 3. The content arrangement change method according to claim 1, wherein when the content causing the access load is specified, a predetermined criterion is used, and if the criterion is exceeded, the content is changed as the cause of the access load. 4. When predicting a time exceeding a limit value of the access load, an access frequency model and a service duration model are configured for each content and each service type from a past access history. The method according to claim 1, wherein the service duration model and the log information are evaluated. 5. The access frequency model for each content tabulates an access occurrence interval based on an access occurrence time, and the access continuation model for each service type calculates an average of the access occurrence intervals of all contents belonging to the service type. The service duration model for each content uses the distribution of how long the service has been displayed at a time rate with respect to the entire display time of the content, and the service duration model for each service type is 5. The content arrangement change method according to claim 4, wherein a distribution obtained by summing values obtained by multiplying a service duration model by the number of accesses for all contents belonging to the service type and dividing by a total number of accesses is used. 6. When predicting a time exceeding a limit value of the access load, for an access that is currently occurring, an access time for continuing a service is estimated from the service duration model, and a newly generated access is generated. 6. The content arrangement change method according to claim 4, wherein the access is estimated based on the service duration model assuming that an access occurs according to the access frequency model, and the load is predicted by integrating them. . 7. The content according to claim 4, wherein a weight is assigned to an access frequency model based on freshness information included in the log information of the content, for a new content or a content having an insufficient number of accesses. How to change the placement. 8. A method for evaluating the log information, the access frequency model, and the service duration model, and issuing a content arrangement change to a media server to be operated by the media server group based on the evaluation. Item 5. The content arrangement changing method according to any one of Items 1 and 4. 9. A plurality of media servers for delivering contents with continuous media typified by video and audio on demand, a plurality of terminal devices accessing the media servers and displaying the contents, A multimedia service system comprising a local communication network for transferring content between the server and the terminal device in real time, or
A content arrangement change system in which a plurality of multimedia service systems are connected via a wide area communication network for transferring contents in real time, wherein the local communication network or each of the media servers connected to the wide area communication network is connected. Access information collecting means for collecting location information, freshness information, and access information from the terminal device for the content to be stored; and a cause of access load when a large number of terminal devices request a content of a specific media server. Load cause content specifying means for specifying content to be used; access information analyzing means for analyzing access information from the terminal device to the content specified by the load cause content specifying means; and an access load that the media server can process. For predicting time beyond the threshold Value predicting means, before the access load of the media server reaches a limit value, a copy of the content, or a copy / relocation for relocating the location between the content and other content.
A content arrangement change system comprising a content arrangement change device having a rearrangement means. 10. The access information collecting means includes means for collecting log information from a group of media servers and extracting from the log information information relating to an access occurrence time, a service continuation time, and a location of a terminal device that has requested access. The content arrangement change system according to claim 9. 11. The content arrangement change system according to claim 9, wherein said load cause content specifying means uses a predetermined criterion to cause an access load when the criterion is exceeded. 12. The limit value predicting means configures an access frequency model and a service duration model for each content and each service type from a past access history, and stores the access frequency model, the service duration model, and the log information. 10. The content arrangement change system according to claim 9, further comprising an evaluation means for evaluating. 13. An access frequency model for each content tabulates an access occurrence interval based on an access occurrence time, and an access continuation model for each service type calculates an average of access occurrence intervals of all contents belonging to the service type. The service duration model for each content uses a distribution of how long the service has been displayed at a time ratio with respect to the entire display time of the content, and the service duration model for each service type is: 13. The content arrangement change system according to claim 12, wherein a distribution obtained by adding values obtained by multiplying a service duration model by the number of accesses for all contents belonging to the service type and dividing by a total number of accesses is used. 14. The limit value predicting means estimates an access time for continuing a service from a service continuation time model for a currently occurring access, and according to an access frequency model for a newly occurring access. And means for estimating an access time from a service duration model assuming the occurrence of the access and predicting a load by integrating them.
How to change the content placement described. 15. The method according to claim 12, further comprising: means for assigning a weight to an access frequency model based on freshness information included in the log information of the content, for new content or content for which the number of accesses is insufficient. Content relocation system. 16. The duplication / relocation means evaluates the log information, the access frequency model, and the service duration model, and changes a content arrangement based on the evaluation as an operation target of the media server group. 13. The content arrangement change system according to claim 9, further comprising means for issuing to the media server. 17. A plurality of media servers for transferring content with continuous media typified by video and audio in real time in response to a request from a terminal device, so that the load is not biased to a specific media server. A storage medium storing a content arrangement change program for performing location information, freshness information, and access information from the terminal device on the content stored in each of the media servers via a local communication network or a wide area communication network. Access information collection process to collect the content of a specific media server, a load cause content identification process of identifying content causing an access load when a request is generated from a large number of terminal devices, and the load cause content identification. Said terminal device to the content identified by the process An access information analysis process for analyzing access information from the server; a limit value prediction process for predicting a time exceeding a limit value of an access load that the media server can process; and before the access load of the media server reaches the limit value, Duplicate or duplicate that creates a copy of the content or rearranges the location of the content and other content
A storage medium storing a content arrangement change program, characterized by having a rearrangement process. 18. The access information collecting process collects log information from a group of media servers and extracts from the log information information relating to an access occurrence time, a service continuation time, and a location of a terminal device that has requested access.
A storage medium storing the content arrangement change program according to claim 7. 19. The storage medium storing the content arrangement change program according to claim 17, wherein the load cause content specifying process uses a predetermined criterion to cause an access load when the criterion is exceeded. 20. The limit value predicting process estimates an access time for continuing a service from a service duration model for a currently occurring access, and follows an access frequency model for a newly occurring access. 18. The storage medium according to claim 17, further comprising a process of estimating an access time from a service duration model assuming that the access has occurred, and estimating a load based on the estimated access time. 21. The method according to claim 20, further comprising the step of assigning a weight to an access frequency model based on freshness information included in the log information of the content, for new content or content with an insufficient number of accesses. A storage medium storing a content arrangement change program. 22. The duplication / relocation process evaluates the log information, the access frequency model, and the service duration model, and changes a content arrangement based on the evaluation as an operation target of the media server group. 21. A storage medium storing a content arrangement change program according to claim 17, which includes a process for issuing to a media server.