JP2002197002A - System and method for autonomously distributed contents delivery - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a system and method for autonomously distributed delivery, by which contents can autonomously be updated and so on. SOLUTION: The autonomously distributed delivery system equipped with a plurality of server devices for distribution which can provide contents data stored in a cache storage means for a client device at a request made by the client device is equipped with a contents providing device as a primitive provision source for the contents data, a cache state controller which manages the cache storage means for one of a plurality of server devices for distribution, and a distribution processing plan implementing device which stores a distribution processing plan regarding the contents data, autonomously controls the operation of the cache state controller according to the distribution processing plan, and transfers the contents data between the server devices for distribution.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to an autonomous distributed content distribution system, for example, a WW on the Internet.
It is suitable to be applied to a case where distribution management of rich contents such as audio and images is executed in a W (World Wide Web: hereinafter abbreviated as Web) environment.

[0002] The present invention also relates to an autonomous distributed content distribution method.

[0003]

2. Description of the Related Art Conventionally, this type of content management method /
The system includes an invention disclosed in the following reference 1 (US patent).

Reference 1: Method and apparatus for dyn
amic cache preloading across a network [Patent Number: 6,098,096]
The present invention relates to a method and a system of a cache server having a function of preloading contents from an original server in a planned manner based on “predetermined criteria” defining criteria such as the size of contents and the relevance between contents.

[0005] According to the method of the present invention, a preloading management agent installed in a cache server is used.
This is to preload content specified in “predetermined criteria” from the original server to the cache memory on the cache server, and to realize content management based on Peer to Peer connection between the original server and the cache server I have.

[0006]

However, in the method and system according to the above-mentioned invention, the cache server that processes the content access request from the client can perform only Peer to Peer communication with the original server. There are the following problems 1) to 3).

(1) In a state where connection with the cache server cannot be established due to a network failure on the original server side, the cache cannot be updated.

(2) If the load on the original server is high or the communication cost with the original server is high, efficient planned distribution cannot be executed.

(3) There is no method for dynamically changing “predetermined criteria”, and it is not possible to flexibly respond to changes in the operating environment.

Furthermore, since a cooperative mechanism between cache servers when a plurality of cache servers are installed is not specified, the following problems (4) and (5) need to be solved.

(4) When the disk capacity on the cache server becomes insufficient, it is necessary to delete the preloaded content from the cache memory. When an access request to the deleted content occurs, communication with the original server is performed. , The access request from the client cannot be sped up.

(5) When a content update occurs, all cache servers need to communicate with the original server and acquire the content from the original server. Therefore, concentration of a network load and a server load occurs.

[0013] These problems (1) to (5) are as follows.
The connection between the original server and the cache server is Peer t
o It can also occur in cases other than Peer.

[0014]

In order to solve such a problem, in an autonomous distributed type distribution system according to the present invention, content data stored in a cache storage unit is transmitted to a client device in response to a request from a client device. In an autonomous distributed distribution system including a plurality of distribution server devices that can be provided to the device, (1) a content providing device that is a primitive provider of the content data; and (2)
A cache state control device that manages a cache storage unit for one or a plurality of the distribution server devices;
(3) Distribution that stores a distribution processing plan related to the content data, autonomously controls the operation of the cache state control device according to the distribution processing plan, and executes the transfer of the content data between the distribution server devices. A processing plan execution device.

Further, in the distributed autonomous distribution system of the present invention, one or a plurality of distribution servers capable of providing content data stored in the cache storage means to the client device in response to a request from the client device. In an autonomous distributed distribution system having a plurality of distribution site systems provided with devices, (1) a content providing device that is a primitive provider of the content data is provided, and (2) each of the distribution site systems is (3) a cache state control device that manages cache storage means for one or a plurality of the distribution server devices;
Storing a distribution processing plan for the content data,
A distribution processing plan execution device that autonomously controls the operation of the cache state control device according to the distribution processing plan and executes the transfer of the content data between the distribution server devices or the distribution site systems. It is characterized by having.

Further, in the distributed autonomous distribution method of the present invention, in response to a request from a client, a plurality of distribution servers provide content data stored in a cache storage means to the client. In the distribution method, (1) when the content providing unit primitively provides the content data, (2) the cache state control unit manages the cache storage means of one or more of the distribution servers, and (3) distribution The processing plan execution unit stores a distribution processing plan related to the content data, controls the operation of the cache state control unit autonomously according to the distribution processing plan, and executes the transfer of the content data between the distribution servers. It is characterized by doing.

Further, in the distributed autonomous distribution method of the present invention, one or a plurality of distribution servers capable of providing content data stored in the cache storage means to the client in response to a request from the client are provided. In the autonomous distributed distribution method in which a plurality of distribution site systems provided cooperate with each other, (1) when the content providing unit primitively provides the content data, (2) in each of the distribution site systems, the cache status A control unit manages a cache storage unit for one or a plurality of the distribution servers, and (3) a distribution processing plan executing unit stores a distribution processing plan relating to the content data, and autonomously operates according to the distribution processing plan. Controlling the operation of the cache state control device, and between the distribution servers or the distribution server. And executes the transfer of the content data between collection system.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION (A) Embodiment One embodiment of an autonomous decentralized content distribution system and method according to the present invention will be described below using a P2P server machine as an example.

In the present embodiment, for the purpose of speeding up content distribution processing in response to a content access request from a client, the present embodiment relates to content management on a plurality of content distribution servers distributed and arranged on a wide area network. It is characterized in that each content distribution server itself autonomously and decentralizedly executes update management of content to be executed and search processing corresponding to a content request from a client.

(A-1) Configuration of the Embodiment FIG. 1 shows a configuration example of a main part of the content distribution management system 10 according to the present embodiment. The content distribution management system 10 is arranged on a WAN (Wide Area Network) 15 such as the Internet.

In FIG. 1, the content distribution management system 10 includes client machines 11A to 11C,
Distribution sites 12A to 12D and content providing machine 1
6A to 16C, a content management instruction mechanism 17, and a wide area service cooperation mechanism 18.

In the distribution site 12A, a plurality of distribution server machines 13A to 13C and a plurality of P2P (PeP)
er to Peer) server machines 14A to 14C are provided. Distribution sites 12B other than distribution site 12A
The internal configuration of 12D is the same as that of the distribution site 12A.

However, reference numerals 13D to 13F are assigned to the distribution server machines in the distribution site 12B, and reference numerals 14D to 14F are assigned to the P2P server machines, so that the distribution server machine and the P2P server machines in the distribution site 12A are Distinguish.

Each of the distribution sites 12A to 12D may logically have a hierarchical structure such as a head office, a branch office, or the like, and physically constitute separate LANs (local area networks). I have.

The content providing machines 16A to 16C
Machines that are managed and operated by different content providers, respectively, are original content servers that are primitive providers of content distributed by the content distribution of the present embodiment. In the case of the present embodiment, the content once provided from the content providing machine (for example, 16A) to any of the distribution servers (for example, 13A) autonomously (that is, the state of the content providing machine (for example, if the content providing machine fails). Irrespective of the occurrence of a load or a concentrated load), the data is transferred between the distribution server machines or deleted on the distribution server machine according to the planned distribution information PD described later.

Various examples of contents to be provided are conceivable, and typical examples include static text / sound / image data / program and dynamic streaming data.

The content management instructing mechanism 17 is an instructing mechanism for the content providing machine (16A or the like) to request the P2P server machine (14A or the like) for planned distribution of the content.

When a content distribution machine (16A or the like) requests a planned distribution of content, the contents of the planned distribution information PD, which specifically indicates the mode of the planned distribution, are mainly content providers who operate the content distribution machine. The distribution site (e.g., 12A) can also participate in the determination or change the content as necessary.

The wide area service cooperation mechanism 18 is connected to a plurality of P2P server machines (for example, one
4A and 14D) are provided. The application of the Jini Lookup service is one of the typical wide-area server linkage mechanisms.

The client machines 11A to 11C are information processing devices operated by a user who uses contents provided by a content providing machine (16A or the like). W
When the web environment is targeted, the client machine may be a personal computer equipped with a web browser.

The distribution server machines (for example, 13A) provided inside each of the above-mentioned distribution sites (for example, 12A) are machines equipped with substantially the same cache function. Or another distribution server machine (eg, 13B, 13B)
D) and the like, the contents that are systematically collected according to the plan distribution information PD are stored in a storage device (for example, 19A) such as a hard disk mounted therein, and a content access request is made from a client machine (such as 11A). When there is a content (for example, CA) corresponding to the content access request from its own storage device (for example, 19A) without accessing the wide area network 15 such as a content providing machine (for example, 16A).
Is read and provided to the client machine.

Note that the content access request is W
In the case of the eb environment, it is issued as an HTTP request.

Since the storage capacity of the storage device (for example, 19A) mounted on the distribution server machine (13A or the like) is limited, the contents (CA
To CX) exceeds the limit, some content (for example, content CC) is deleted from the stored content, and new content is cached. Thus, a content access request may be supplied from a client machine in order to provide the deleted content (CC).

Therefore, the process performed by the distribution server machine (13A or the like) is that if the content requested by the content access request is stored in the storage device of the distribution server machine that has received the content access request, the content is immediately The process is terminated only by distributing to the client machine, but if the data is not stored, it is necessary to cooperate with a P2P server machine (for example, 14A) according to a procedure described later.

Each of the distribution server machines may be, for example, a proxy server or a Web server compatible with the WWW (World Wide Web).

Each P2P server machine (14) that manages one or more distribution server machines (13A, etc.)
A to 14F) have substantially the same function,
Each of them has predetermined distribution information IS regarding the distribution server machine to be managed.

Here, the P2P server machine 14A manages the distribution server machine 13A, and the P2P server machine 1A
4B manages the distribution server machine 13B, the P2P server machine 14C manages the distribution server machine 13C,
The P server machine 14D manages the distribution server machine 13D, and the P2P server machine 14E
E is managed, and the P2P server machine 14F manages the distribution server machine 13F.

A configuration example of the distribution information IS may be as shown in FIG. Each P2P server machine 14A-14
The distribution information IS held by F has substantially the same configuration, but FIG. 2 is described as showing the ISA which is the distribution information IS related to a certain content held by the P2P server machine 14A. Advance.

In FIG. 2, the distribution information ISA includes a cache target content identifier TC and the plan distribution information P
D, access frequency AP, distribution cost PC between distribution servers, and logical configuration information LC.

Here, the content identifier TC to be cached uniquely specifies the content to be held and managed as cache data on the distribution server machine (for example, 13A) (ie, on a storage device mounted on the distribution server machine). An identifier.

Other information PD and AP on the distribution information IS
Is also information about the content specified by the cache target content identifier TC.

That is, the plan distribution information PD indicates the specific contents of the distribution plan, and the access frequency AP indicates the access frequency (from each client machine) almost in real time.

The planned distribution information PD is mainly designated by a content provider who operates a content providing machine (16A or the like) in advance when the content is provided, and is a distribution server machine that distributes the content. (Such as 13A) (spatial designation),
Designation of the distribution time to the distribution server machine (time specification), specification of the distribution server machine from which the content is deleted (spatial specification), and specification of the deletion time to delete the content from the distribution server machine (temporal specification) Specified).

The access frequency AP is statistical information indicating the frequency of access to the content from the client machine. The access frequency AP is the total number of accesses from the time when a certain content is distributed to a distribution server device (for example, 13A) to the present time. Or the instantaneous value of the access (for example,
(The maximum number of times of access in time).

When there are a plurality of distribution server machines that cache the same content, the access frequency AP can be different for each distribution server machine (for example, 13A and 13B), and thus is dynamically changed for each distribution server machine. .

Further, when there are a plurality of contents (for example, CA to CX) cached by the same distribution server machine (for example, 13A), the access frequency AP is different for each content. When it is necessary to delete the content, in order to determine which content is to be deleted (this deletion is exceptionally performed against the planned distribution information PD of the content), the access frequency AP is a determination criterion. obtain.

On the contrary, the distribution cost PC and the logical configuration information L
C is information that does not depend on the content.

That is, the distribution cost PC stores cost information such as the width of a communication band necessary for transferring contents between the distribution server machines, the communication time, and the communication distance to a set of distribution server machines (for example, 13A and 13D). It is shown for each. Generally, it is considered that as the value of the distribution cost PC becomes wider and longer, the usage fee of the communication line provided by the communication carrier becomes higher.

The logical configuration information LC in the distribution information IS
Is a data structure such as a list structure or a tree structure (also included in a broad list structure), but each P2P server machine (eg, 14A)
However, since it is not necessary to know the logical configuration of all the distribution sites (for example, 12A to 12D) and the logical configuration of all the distribution servers (13A to 13F, etc.) existing in the content distribution management system 10, Only the data of the part needs to be held as the logical configuration information LC.

One P2P server machine (for example, 1
When 4A) searches for the target content in the content distribution management system 10, it follows each P2P server machine (14B, 14D, etc.) in order according to the logical configuration.

Specific examples of the logical configuration include a hierarchical configuration (tree structure) representing a parent-child relationship, as well as a group configuration representing an equal relationship.

FIG. 3 shows an example of the internal configuration of the P2P server machines 14A to 14F. P2P server machines 14A-14
Since F has substantially the same internal configuration, the description will proceed mainly with FIG. 3 showing the internal configuration of the P2P server machine 14A.

(A-1-1) Internal Configuration of P2P Server Machine In FIG. 3, the P2P server machine 14A includes a content search mechanism 30, a content update mechanism 31, a content directory management mechanism 32, and various service mechanisms 33. , A content transfer mechanism 34.

In the case where the content requested by the client machine does not exist in the distribution server machine (for example, 13A) managed by the content search mechanism 30, the content search mechanism 30 first operates the self-delivery site (for the P2P server machine 14A). Is a portion for executing a content search in the distribution site 12A) and requesting a content search outside the own distribution site when the target content cannot be searched by the search.

First, a content search is executed in the self-distribution site, and a search outside the self-distribution site is executed only when the target content cannot be searched. Usually, the content is transferred in the self-distribution site. This is because the distribution cost PC is lower.

Each P2P server can recognize only the content of the content stored in the distribution server machine only from the distribution server machine managed by the P2P server. Also, even if the search is performed outside the own distribution site, the contents stored in the distribution server machines other than the distribution server machine (here, only 13A) managed by the own P2P server machine (here, 14A) are stored. When searching, the corresponding P2
It is necessary to issue a search request to the P server machine.

Note that the content search is not necessarily performed for all P2P server machines in the content distribution management system 10, and ends when the target content can be searched.

The search by the content search mechanism 30 may be performed in order to store the target content in the distribution server machine 13A managed by itself, but other P2P
In some cases, a search is requested from a server machine (for example, 14B). However, the requested search is executed only for the distribution server machine 13A managed by itself.

The content update mechanism 31 is a distribution server machine 1 managed by its own P2P server machine (14A).
When the content search mechanism 30 executes a search or issues a search request to store the target content in 3A, and the target content can be searched, the planned content is updated according to the plan distribution information PD. (Addition, deletion, change).

The content update mechanism 31 also changes the distribution information IS shown in FIG.
, A function of autonomously issuing an update request to a P2P server on another distribution site is provided.

Therefore, after a certain content is primitively provided from a content providing machine (for example, 16A), the content is stored in at least one of the delivery server machines in the content delivery management system 10, and the content related to the content is provided. As long as the distribution information IS is held in the corresponding P2P server machine, it is possible to subsequently update the contents of each distribution server machine in the content distribution management system 10 in a chained manner.

Since the necessity of updating the content can be recognized only by examining the plan distribution information PD, the content update is performed by the P which stores the plan distribution information PD.
Depending on the relationship between the 2P server machine and the contents of the plan distribution information PD, it can be classified into active update and inactive update.

A P2P server machine (for example, 14A)
Actively updating the content by recognizing the necessity of updating the content with respect to the distribution server machine (for example, 13A) managed by itself is active updating, and the other P2P server machine ( For example, 14D)
Recognizes necessity and executes content update
The case where necessary content is transferred to the server machine is an inactive update.

In the case of active update, each P2P server machine needs to execute a content search for required content prior to the content update in order to update the content of the distribution server machine managed by itself. There is.

On the other hand, in the case of an inactive update, another P2P server machine recognizing the necessity of the content update executes a content search and transfers the content related to the content update. Therefore, the P2P server machine that updates the content need only accept the transferred content.

However, depending on the system configuration, other P2
When the P server machine (for example, 14D) recognizes the necessity of the content update, it notifies only the necessity, and the content search itself is performed by the P2P server machine (for example, 14A) that receives the notification and executes the content update. It may be executed.

The contents directory management unit 32
Is a management mechanism that manages a directory of contents held in the distribution server machine 14A managed by the own P2P server machine 14A.

The content directory allows the content identifier TC to be cached in the distribution information IS.
Can be referred to.

The various service mechanisms 33 are provided with various functions associated with a content management function such as an access control function based on authentication of a client machine that issues an access request to a content or a user who operates the client machine, and collection of an access log. A mechanism for supporting services.

Finally, the content transfer mechanism 34 sets its own P
Delivery server machine 1 managed by 2P server machine 14A
For the purpose of updating (in this case, adding or changing) the content of 3A, the content is transferred to the storage device 19A to store new content, or the content is transferred to another P2P server machine. In order to update (for example, add or change) the content, the content read from the storage device 19A is transferred to a distribution server machine (for example, 13B) managed by another P2P server machine. ,
This is a mechanism that handles content transfer.

The operation of the present embodiment having the above configuration will be described below with reference to FIGS.

(A-2) Operation of the Embodiment Here, the distribution server machine 13A in FIG. 1 and the P2P server machine 14A for managing it, the distribution server machine 13D and the P2P server machine 14D for managing it,
A description will be given focusing on the operation between the distribution server machine 13E and the P2P server machine 14E that manages the distribution server machine 13E.

In FIG. 4, each P2P server 14D, 1
4E starts up with the P necessary to access itself.
The 2P server proxies PRD and PRE and the distribution information ISD and ISE stored therein are registered in the wide area service cooperation mechanism 18.

Here, the P2P server proxy PRD is connected to another P2P server machine (here, the P2P server machine 1).
4A) is an interface for communicating with the P2P server machine 14D. By using such a proxy, when communicating with the P2P server machine 14D, the P2P server machine 14A is released from the burden of handling information such as the detailed address of the P2P server machine 14D, and uses a simple interface. The communication can be executed by simple processing.

Similarly, the P2P server proxy PRE is connected to another P2P server machine (here, P2P server machine 1).
4A) is an interface for communicating with the P2P server machine 14E.

In the wide area service cooperation mechanism 18, P
2P server machine 14A to P2P server machine 14D
When requested to start communication with the P2P server machine 1, the P2P server proxy PRD and the distribution information ISD are
Transfer to 4A.

Similarly, the wide area service coordinating mechanism 18
When requested to start communication with the P2P server machine 14E from the P server machine 14A, the P2P server proxy P
The RE and the distribution information ISE are transferred to the P2P server machine 14.
Transfer to A.

P2P server proxies PRD, PRE,
The P2P server machine 14A that has received the transfer of the distribution information ISD and ISE communicates with the P2P server machine 14D by using the P2P server proxy PRD, and communicates with the P2P server machine 14E by using the P2P server proxy PRE.
Communicate with

In FIG. 4, the P2P server machines 14A and P
The communication of the 2P server machine 14D is shown as service A, and the P2P server machine 14A and the P2P server machine 1
The communication of 4E is shown as service B. Service A
And service B may be the same type of communication or different types of communication. For example, the content search and the content update described above are realized by the communication, and the execution log information collection service held and managed by the various service mechanisms 33 of the P2P server machine (for example, 14D) as the communication partner, and the client. A service related to access control by authentication may be realized by the communication.

When a content search is performed, as shown in FIG. 5, a search program 3 corresponding to the content search mechanism 31 of the P2P server machine 14A is executed.
0A uses the P2P server proxy PRD, PRE,
A search request for a content search is issued to each of the P2P server machines 14D and 14E that are communication partners, and communication is performed.

The content search shown in FIG. 4 and FIG. 5 is performed between P2P server machines having different distribution sites, but such a content search is executed next to the following content search. Things.

That is, the P2P server machine 14A is connected to an external (other than its own P2P server machine) P2P
When issuing a search request to a server machine, prior to the issuance, first, the distribution server machine 13 managed by itself.
It is necessary to search for the content stored in the storage device 19A of A. If the target content cannot be searched by this search, the other P2P server machines 14B and 14C existing in the own distribution site 12A manage the content. When a search request is issued to the distribution server machines 13B and 13C, and the target content cannot be searched by the search, a search request is issued to the P2P server machine (14D or the like) belonging to a different distribution site. .

Here, the issuance of a search request to another P2P server machine (for example, 14B) in its own distribution site is as follows.
4 and 5, but the same LA
Since the processing is performed within N (distribution site), it is considered that a more efficient and faster method can be used according to the specification of the LAN.

The search for the content stored in the storage device 19A of the distribution server machine 13A managed by itself is executed according to the internal specifications of the distribution server machine 13A and the P2P server machine 14A. Become.

The search may be performed by the P2P server machine 14A, but the distribution server machine 13A itself may be provided with a function for performing the search.

When the content access request is supplied to the distribution server machine 13A, the corresponding P2P server machine 14A stores the content requested by the content access request in the storage device 1 of the distribution server machine 13A.
If the content is a content (for example, CA) stored in 9A, the content is immediately read from the storage device 19A in the distribution server machine 13A and provided to the client machine 11A.

In the case of this embodiment, the P2P server machine 1
4A also receives the transfer of the distribution information ISD, ISE together with the P2P server proxies PRD, PRE, and thus preliminarily executes a certain degree of content search (execution of preliminary search) based on the distribution information ISD, ISE received by this transfer. It is possible. Further, by analyzing the plan distribution information PD in the distribution information (for example, ISD), it is possible to predict the distribution server machine storing the target content at present at a certain degree of accuracy (predicting the position of the content). is there.

However, the distribution information (for example, ISD) registered in the wide area service cooperation mechanism 18 is a P2P server machine (for example, 14D) which is likely to exist in a remote place.
Is registered at the time of activation of the server, and does not necessarily provide real-time information of the corresponding distribution server machine (for example, 13D).

Further, it is difficult to predict in advance whether or not a certain content is stored in an actual distribution server machine (for example, 13D) due to dynamic factors. For example, if the content having the planned distribution information PD desired to be distributed to the distribution server machine is concentrated on a specific distribution server machine, the storage capacity of the storage device mounted on the distribution server machine becomes insufficient, and all of the plans It may be impossible to faithfully perform the distribution information PD.

Therefore, if the preliminary search and the prediction of the content position are utilized to determine the search priority when selecting the P2P server machine to be searched in the content search, the content search and the content are predicted. While it ’s likely to be the most efficient update, in order to ultimately achieve content search and content updates,
It is necessary to actually communicate with a P2P server machine as a communication partner like the service A or the service B.

Each P2P server machine (for example, 14
The distribution information (for example, ISD) registered by D) in the wide area service cooperation mechanism 18 includes the information elements TC, P shown in FIG.
With D, AP, PC, LC, or in place of these information elements, operation state information indicating the operation state of the distribution server machine (for example, 13D) managed by the P2P server machine that has registered the distribution information (load concentration degree) Or information indicating whether a failure has occurred or the like).

On the other hand, a content access request inconsistent with the planned distribution information PD, that is, a content that is requested for the content during a period in which the content is not distributed to the distribution server machine (for example, 13A) according to the planned distribution information PD. When the access request is supplied from the client machine (for example, 11A),
Various things can be considered.

For example, according to the planned distribution information PD, at this time, a message indicating that the distribution of the content is not performed from the distribution server machine (for example, 13A) that has received the content access request (in this message, It is also possible to adopt a configuration in which information for notifying a distribution server machine (for example, 13B) predicted that the content is currently distributed is returned to the client machine. is there.

However, in the present embodiment, even in such a case, in order to respond to the request of the client machine,
A corresponding content search and content update are performed.

The content search process itself is the same as the content search executed based on the plan distribution information PD. The content update process itself is the same as the content update performed based on the plan distribution information PD.

Next, FIG. 7 shows a state of the content update operation executed based on the plan distribution information PD.

In FIG. 7, the contents (for example, the C
When B) is supplied from the content providing machine (for example, 16A) to the content management instruction mechanism 17, the planned distribution information P based on the designation from the content providing machine 16A or the like is obtained.
D, the content CB is, for example, distributed site 1
It is stored in the distribution server machine 13A in 2A.

Thereafter, each time the content is updated, the content CB constitutes an update unit UC together with at least its own plan execution program PD1P (including the cache target content identifier TC, etc.). The data is transferred to another set of the distribution server machine and the P2P server machine in units of UC.

Here, the plan execution program PD1P is
This is a program for executing the contents of the plan distribution information PD.

When the content is updated, the distribution server machine at the transfer destination of the update unit UC naturally stores the content, but the transfer server machine at the transfer source maintains or deletes the storage of the content CB. This is determined according to the contents of the plan distribution information PD.

In FIG. 7, each P2P server machine 14A,
14D, PC, and PD recursively issue a content update request (that is, transfer the update unit UC) in accordance with their own logical configuration information LC.

FIG. 7 shows a case where the distribution site 12A is the head office, and the distribution sites 12B and 12C are the branch offices of 12A.
The distribution site 12D corresponds to a case where the distribution site 12D is a subsidiary of the branch office 12B.

In order for the content distribution management system 10 of the present embodiment to autonomously repeat content distribution independently of the content providing machine (16A or the like), the content once provided from the content providing machine (for example, CB ) Requires that the content CB be stored in any of the distribution server machines (13A to 13F, etc.) in the content distribution management system 10 until the plan distribution information PD reaches the end of the plan. .

This means that, for example, in any of the distribution server machines (for example, 13A), the storage device (19A)
Of content (for example, CB)
This is a problem when it is necessary to continue distribution according to the planned distribution information PD, but it must be deleted.

In such a case, as shown in FIG.
The content CB may be temporarily saved (relocated) to another distribution server machine (for example, 13D) having a sufficient storage capacity.

Whether or not there is enough storage capacity can be determined based on the distribution information (ISD, etc.) shown in FIG. 4, or can be explicitly confirmed by communication between P2P server machines. It is possible.

It is sufficient to perform such a save for one distribution server machine. Therefore, when there are a plurality of distribution server machines having a sufficient storage capacity, the distribution server with the lower distribution cost is used. A server machine (for example, 13D in FIG. 6) may be selected as a save destination.

In the present embodiment as described above, the wide area service cooperation mechanism 18 can mutually connect the distribution sites distributed on the wide area network.

Then, it is possible to execute content management based on mutual communication between the connected distribution sites.

Further, it is possible to recognize the operating status and network fault status of the connected distribution site, and to dynamically change the content management method according to the status.

Further, based on the distribution information (such as the ISD) managed by the P2P server, an appropriate distribution site with which the own P2P server communicates can be limited, and efficient content management can be realized. .

Further, when an access request from a client to the content deleted from the cache occurs due to the movement of the content between the distribution server machines, a distribution site having a low distribution cost is obtained instead of being obtained from the content providing machine. By issuing the content search to (the P2P server machine among them), it is possible to efficiently respond to the access request.

Further, the autonomous content update process between the P2P server machines can reduce the direct communication process with the content providing machine.
Network traffic can be distributed.

(A-3) Effects of Embodiment As described above, according to the present embodiment, it is possible to realize autonomous and decentralized content management (content distribution) in cooperation with each distribution server machine. Become.

Thus, for example, a content providing machine (or a LAN including the content providing machine)
Even if a failure occurs on the side, the load is concentrated, or the communication cost is high, the content distribution according to the planned distribution information can be continued efficiently, and the reliability is improved.

Since the access frequency (AP) can be dynamically changed, it is possible to flexibly respond to a change in the operating environment.

Further, in the present embodiment, since each distribution server machine can be favorably linked, if the storage capacity of any of the distribution server machines is insufficient, the contents which had to be deleted in the past can be deleted. Can be evacuated to the distribution server machine, so the utilization efficiency of hardware resources is high and the reliability is high.

Furthermore, in the present embodiment, even when the content is updated, the load is not concentrated on the original server (content providing machine) as in the related art, and the network load and the server load can be efficiently distributed. .

(B) Other Embodiments In the above embodiment, the distribution information notified to the wide area service cooperation mechanism does not include the information of the content held by each distribution server (that is, the distribution status information) at that time. However, it is also possible to include distribution status information. In the case of this configuration, the search processing over a wide area from the P2P server machine is unnecessary, and the search processing can be completed only inside a specific P2P server. Content can also be obtained from server machines. This can be expected as an effective method when implementing content management using a small operating site.

In the above embodiment, the distribution server machine (13A) in one distribution site (for example, 12A)
~ 13C) and P2P server machine (14A ~ 14C)
Are the same, but these numbers may be different. Considering that a single P2P server machine often has a configuration in which a plurality of distribution server machines are managed, P
Usually, the number of distribution server machines is larger than the number of 2P server machines.

Although the number of the client machines 11A to 11C and the number of the content providing machines 16A to 16C shown in FIG. 1 of the above embodiment are the same, they do not need to be the same, and the three shown in FIG. It may be more or less than one.

Further, in the above description, the component 14A
14F and the like are P2P server machines, but in the present invention, the components can be replaced with information processing apparatuses having communication functions other than P2P.

In the above embodiment, the logical configuration information LC
Has nothing to do with the content, but the logical configuration information may be set according to the content.

Further, in the above embodiment, each component (for example, 19A, 14A, 16A, etc.) has been mainly described as hardware, but these may be realized by software having equivalent functions. That is natural.

[0125]

As described above, according to the present invention, after the content data is primitively provided from the content providing device (or the content providing unit), each delivery server device (or the content data) of the content data is provided. The distribution to each distribution server) is executed autonomously according to the distribution processing plan, so that the distribution is autonomous and highly reliable.

[Brief description of the drawings]

FIG. 1 is a schematic diagram illustrating a configuration example of a content distribution management system according to an embodiment.

FIG. 2 is a schematic diagram illustrating a configuration example of distribution information used in the embodiment.

FIG. 3 is a schematic diagram illustrating a configuration example of a P2P server machine used in the embodiment.

FIG. 4 is an operation explanatory diagram of the embodiment.

FIG. 5 is an operation explanatory diagram of the embodiment.

FIG. 6 is an operation explanatory diagram of the embodiment.

FIG. 7 is an operation explanatory diagram of the embodiment.

[Explanation of symbols]

10. Content distribution management system, 11A to 11C ...
Client machine, 12A to 12D ... distribution site, 1
3A to 13F: distribution server machine, 14A to 14F: P
2P server machine, 16A to 16C: content providing machine, 17: content management instruction mechanism, 18: wide area service cooperation mechanism.

 ────────────────────────────────────────────────── ─── Continued on the front page (72) Inventor Kazuhiro Kawabe 1-7-12 Toranomon, Minato-ku, Tokyo Oki Electric Industry Co., Ltd. F-term (reference) 5B082 FA12 HA05 5B085 BA06 BC01 BG07 CA04 CA07

Claims (9)

[Claims] 1. In response to a request from a client device,
In an autonomous distributed distribution system including a plurality of distribution server devices that can provide content data stored in a cache storage unit to the client device, a content providing device that is a primitive provider of the content data; A cache status control device that manages a cache storage means for one or more of the distribution server devices, and a distribution processing plan for the content data,
A distribution processing plan execution device that autonomously controls the operation of the cache state control device according to the distribution processing plan and executes the transfer of the content data between the distribution server devices. Type distribution system. 2. The autonomous distributed distribution system according to claim 1, wherein the information elements constituting the distribution processing plan include: a content identifier that specifies content data to be stored at least temporarily in the cache storage device; And a distribution processing information describing a temporal and spatial distribution plan of the content data specified by the content identifier. 3. The autonomous decentralized distribution system according to claim 2, wherein, as information elements constituting the distribution processing plan, cost information relating to a consumed time or a consumed bandwidth for transferring content data between the respective distribution server devices. An autonomous decentralized distribution system comprising: 4. The autonomous distributed distribution system according to claim 2, wherein the information elements constituting the distribution processing plan include logical information describing a logical structure between distribution server devices. Distributed distribution system. 5. The autonomous distributed type distribution system according to claim 4, wherein the distribution processing plan execution device autonomously executes the distribution processing plan in the autonomous distributed type distribution system, so that the logical information is also stored. An autonomous decentralized distribution system characterized by executing a search for content data in search of target content data. 6. The distributed autonomous distribution system according to claim 5, wherein the distribution processing plan execution device transfers the target content data obtained by the search between the distribution server devices, and An autonomous decentralized distribution system that updates content data by writing the received content data to a cache storage unit in a distribution server device specified by a distribution processing plan. 7. In response to a request from a client device,
In an autonomous distributed distribution system having a plurality of distribution site systems including one or a plurality of distribution server devices capable of providing content data stored in a cache storage means to the client device, Each of the distribution site systems includes a cache state control device that manages a cache storage unit of one or a plurality of the distribution server devices; Store delivery processing plan,
A distribution processing plan execution device that autonomously controls the operation of the cache state control device according to the distribution processing plan and executes the transfer of the content data between the distribution server devices or the distribution site systems. A distributed autonomous distribution system characterized by comprising: 8. An autonomous distributed delivery method in which a plurality of delivery servers provide content data stored in a cache storage means to the client in response to a request from the client, wherein the content providing unit includes the content providing unit. When the data is provided primitively, the cache state control unit manages the cache storage means of one or more of the distribution servers, and the distribution processing plan execution unit stores the distribution processing plan related to the content data, and An autonomous distributed distribution method characterized by autonomously controlling the operation of the cache state control unit according to a processing plan, and executing the transfer of the content data between distribution servers. 9. A plurality of distribution site systems including one or a plurality of distribution servers capable of providing content data stored in a cache storage unit to the client in response to a request from the client. In the distributed autonomous distribution method, when the content providing unit primitively provides the content data, in each of the distribution site systems, the cache state control unit stores the cache storage for one or a plurality of the distribution servers. Means for managing the means, a distribution processing plan execution unit stores a distribution processing plan relating to the content data, autonomously controls the operation of the cache state control device according to the distribution processing plan, and controls the operation between the distribution servers or Executing the transfer of the content data between the distribution site systems. Decentralized delivery method and butterflies.