JP2002318720A - Contents delivery management system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily access target contents. SOLUTION: In the contents delivery managing system for providing uniform contents on a network distributed over a wide area, this system is provided with a wide area load distributing device 1-2 for selecting a low-order site while considering the load of the low-order site or network or the like, P2P servers 1-5 or the like, which are mutually connected peer to peer, for managing the contents or the like while being distributed, and an index server 1-7 for managing all contents information. Further, the system is provided with an index proxy 1-6 for caching the contents information, complementing the index server 1-7 and avoiding the concentration of loads to the index server 107 and a look-up server 1-8 for discriminating whether the server is under staring or not or whether a new server is registered or not and reporting the result to the index server 1-7.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to updating, adding or deleting contents, adding or deleting servers of various sizes, stopping or restarting a server, while allowing different contents to exist in a plurality of contents distribution servers. The present invention relates to a content distribution management system which enables a user to easily search for necessary content.

[0002]

2. Description of the Related Art A conventional content distribution management system has the invention disclosed in the following US Patent.

[0003] 1. System and method
for server-sideoptimizat
ion of data delivery on a
distributed computer network
orkPatent Number: 6,112,23
9 2. Method and system for d
directing aflow between a
client and a serverPatent
Number: 6,006,264 The first invention is an invention relating to a method / system for analyzing network performance between a client and a plurality of content distribution servers and distributing Web content access from the client to an appropriate content distribution server. . In this method, an appropriate content distribution server is selected for each client when a content request is issued from a client based on the contents of a database storing analysis results of network performance.

The second aspect of the present invention relates to a content type / QoS request designation specified in a content request issued from a client, a load of a content distribution server group available at the time of request / network congestion degree / client and server. The distance of the Web site from the client
The present invention relates to a system / system for distributing content access to an appropriate content distribution server. This is a method of managing the collected information as the performance metrics of the content distribution server and selecting an appropriate content distribution server for each client when a content request is issued from the client.

[0005] These patents all have a plurality of servers for distributing content, but it is premised that those servers have the same content.

There are the following conventional file supply systems.

[0007] 1. http: // www. oceans
tore. cs. berkley. edu 2. http: // www. napster. com The first system is a system having a configuration in which a plurality of servers connected to a network are connected in a network-like manner (a server is sparsely connected to all other servers). In this system, it is possible to search for a target file by sequentially tracing servers connected via a network. If you find the file you want, transfer it directly. That is, the file can be directly transferred and acquired from the server having the file, instead of being bucket-relayed according to the search procedure. In this system, the Bloom method is used as search means. There is no server that centrally manages information on files that exist in the system.

In the second system, there is a server for centrally managing information on files existing in the system, and a server having a file must be connected to a server for centrally managing file information. When a server having a file acquires a specific file, the server that centrally manages the file information searches for the specific file, and acquires the specific file directly from another server having the specific file.

[0009]

In the conventional content distribution system, contents (text, audio, video, etc.)
Must be the same for all servers that deliver the content. Therefore, when this system is applied when contents are frequently changed, there is a problem that the more servers that distribute the contents in a wide area, the more difficult it is to equalize the contents in the servers. .

[0010] Further, since all servers for distributing contents require a storage medium such as a large-capacity hard disk capable of storing all contents, it has been difficult to add a small-scale server.

In a conventional file sharing system, a file having the same file name but different contents may be registered, and a target file may not be taken out.

[0012]

According to a first aspect of the present invention, there is provided a content distribution management system for providing uniform content over a wide area distributed network. A wide-area load distribution device that selects a lower-level site in consideration of the network load, etc., and a content distribution that is provided in each of the above sites and that distributes and manages the content and the like and that is interconnected by peer-to-peer. Server and
And an index server which is at the highest level and manages all content information in the system.

With the above configuration, the wide area load distribution apparatus does not recognize the location of the target content, and simply selects the lower site in consideration of the lower site, the load on the network, and the like. The content distribution server of the selected site searches for the target content by referring to the content information managed by itself or the index server. If the target content is on another content distribution server, the content distribution server of the selected site issues a content acquisition request to the content distribution server with the target content, exchanges the content between the content distribution servers, and distributes the content over a wide area. Content that complements each other. Thus, the target content can be easily accessed. When updating, adding, or deleting content in the content distribution server, the content is directly updated, added, or deleted with respect to an arbitrary content distribution server, and the content information is recorded in the index server. As a result, the newly added content can be shared by all content distribution servers, and the deleted content can be recognized by all content distribution servers.

A content distribution management system according to a second aspect of the present invention is the content distribution management system according to the first aspect of the present invention, wherein the content information is cached to supplement the index server, thereby avoiding load concentration on the index server. It is characterized by comprising an index proxy.

With the above configuration, the content information is cached in the index proxy. This content information is a part of the content information recorded on the index server and complements the index server. This avoids load concentration on the index server.

A content distribution management system according to a third aspect of the present invention is the content distribution management system according to the first or second aspect, wherein each server in the system determines whether or not each server is currently activated and is not activated. Is provided with a lookup server that deletes a new server from the list and adds the new server to the list when the new server is registered, and notifies the index server of the addition.

[0017] With the above configuration, the lookup server:
It is determined whether or not each server is currently running, and if there is a server that is not running, it is deleted from the list, and the information is notified to the index server. With this notification, the content information is deleted from the recording. This prevents a server that is not currently running from being searched. If a new server is registered, it will be listed. As a result, the newly added server is searched in the same manner as the existing server, and a search result according to the current state can be obtained.

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A content distribution management system according to the present invention connects content distribution servers scattered at bases distributed over a wide area by peer-to-peer, and distributes contents over a wide area to complement each other. Fitting reduces the load on each server, and also enables dynamic addition / deletion of content, servers, etc., thereby reducing the concentration of load and enabling easy access to the target content, thereby improving the communication quality of the entire system. It is intended to improve the quality. Not all the contents are stored on the hard disks of the Web servers of all sites, but only a part of the entire contents is stored in each P2P server. The required content is stored in any of the P2P servers in the system. That is, a P2P server (content distribution server) connected by peer-to-peer complements contents distributed over a wide area. In addition, since all contents are not held by one P2P server but complement each other, dynamic addition and deletion of contents and sites can be easily performed. As a result, the concentration of data, traffic, and the like is eliminated, and the target content can be easily accessed to improve the communication quality of the entire system.

FIG. 1 is an overall configuration diagram showing a content distribution management system according to an embodiment of the present invention. The outline of each component is shown below.

The content distribution management system mainly includes a client 1-1, a wide area load distribution device 1-2, a site 1-3, a Web server 1-4, and a P2P server 1-5.
, An index proxy 1-6, an index server 1-7, and a lookup server 1-8.

[0021] The client 1-1 is a user who accesses by a Web browser.

The wide area load distribution device 1-2 is a device for distributing the load and selecting a lower site. When a server is designated by a Web browser, the server is designated by a domain name. The wide area load distribution device 1-2 has a function of a domain name server, and selects a lower site in consideration of a lower site, a network load, and the like.

Sites 1-3 are sites scattered over a wide area on the network. In FIG. 1, three sites A,
B and C are described. All sites are connected online via the Internet or the like.

The Web server 1-4 is a Web server capable of calling external programs such as Servlet and CGI. The Web server 1-4 is a general one and does not perform any special operation for the content distribution management system. Also, a plurality of Web servers 1-4 are allowed in the site.

The P2P server 1-5 is a server that manages original contents, contents information and their caches. Each P2P server is connected to each other by peer-to-peer, so that information such as contents can be freely exchanged between the P2P servers. Each P2P
Different contents and the like are appropriately recorded on the server.
Information such as contents stored in the content distribution management system is recorded and distributed somewhere in each P2P server. That is, information such as contents is distributed and managed in a wide area, and complements each other. Web server 1
-4 is connected to P2P server 1-5 by Servlet
This is performed by an external call of the Web server 1-4 such as a Web server or CGI. The content information is information attached to the content, such as a path name in which the content is stored, a content name, a creation date, and a size.

Each P2P server is interconnected by peer-to-peer. As a result, the contents and the like are exchanged between the P2P servers, and the contents distributed over a wide area are mutually complemented. Further, each P2P server exchanges information such as updates with the index proxy 1-6 and the index server 1-7 in accordance with a change, deletion, or addition of content or the like.

The index proxy 1-6 is a server that caches content information. The index proxy 1-6 normally assists the index server 1-7 by placing it in a place where the distance between the sites is short, thereby avoiding concentration of the load on the index server 1-7. If the load is concentrated in other places, the load can be avoided by placing the index proxy 1-6. The index proxy 1-6 does not necessarily need to belong to the site, but only needs to exist on the net.

The index server 1-7 is a server that exists at the highest level and manages all content information in the system. There is only one in this content distribution management system. Also, this index server 1-
7 does not necessarily need to belong to the site, but only needs to exist on the net.

In the lookup server 1-8, a P2P server (1-5, 1-9, 1-10, etc.), an index proxy (1-6, etc.), and an index server (1-7, etc.) are registered. . By using the lookup server 1-8, it is possible to search for a registered server and determine whether the server is currently running. The lookup server 1-8 has a mechanism for checking whether a registered server (index server, index proxy, P2P server) is operating at regular time intervals. If it is determined by this check that it is not operating, it is deleted from the list. As a result, the server registered in the lookup server 1-8 is not perfect, but is likely to be operating. Further, the lookup server 1-8 has a function of notifying that a new server is registered or that the server is not operating. The index server 1-7 has a program for listening to a notification of a status change of a server registered in the lookup server 1-8. Therefore, the index server 1-7 can determine the currently functioning index proxy or P2P server. If the index server 1-7 determines that any of the P2P servers is down, the index server 1-7
7 deletes the original content owned by the P2P server and the content information of the content cache from the index server 1-7. By performing this operation, the content of the down P2P server is excluded from the search target, and it is possible to return the search result according to the current state.

FIG. 2 shows a P2P server (1-5, 1-9, 1).
It is a block diagram of -10). The outline of the P2P server is shown below.

The P2P server mainly includes a distributed environment 2-1 and
P2P module 2-2 and Admin module 2-
3, a database 2-4, and a hard disk 2-5.

The distributed environment 2-1 is a basic environment in which a P2P server operates, and is usually constituted by Jini.

The P2P module 2-2 is a mounting section of an interface for connecting to a P2P server, an index proxy, an index server, and an interface for changing settings of the P2P server.

The Admin module 2-3 is a mounting unit of an interface for performing management. The P2P server and the index proxy have an upper site, and manage the setting of the upper site, the termination of the P2P server, and the like.

The database 2-4 includes a P2P server 1-
5 is a database for storing the content information registered in the storage unit 5, and usually a relational database is used.

The hard disk 2-5 has a P2P server 1
This is an area for storing contents registered in -5.

FIG. 3 is a configuration diagram of the index proxy 1-6 and the index server 1-7. The index proxy 1-6 and the index server 1-7 have the same configuration. The index proxy 1-6 and the index server 1-7 mainly use the distributed environment 3-1 and P
2P module 3-2 and Admin module 3-3
And a database 3-4. Each component is the same as the P2P server shown in FIG. In addition,
Index proxy 1-6 and index server 1
In -7, there is no hard disk 2-5 which is an area for storing contents.

The content distribution management system configured as described above operates as follows.

[First Operation] A case where a specific HTML file (content) is obtained by a Web browser will be described with reference to FIG. FIG. 4 is an access sequence showing a flow of acquiring content.

(1) The client 1-1 inputs a URL using a Web browser or the like. The client issues a request for searching for an IP address to the wide area load distribution device 1-2 as a domain name server.

(2) Since the wide area load distribution device 1-2 operates as a server that manages the domain, it receives the information. The wide area load distribution device 1-2, based on the information, loads the Web server, the network,
In consideration of a network distance to the client 1-1, an appropriate server is selected from a plurality of lower-level Web servers, and an IP address is returned. Here, a case will be described where the Web server 1-4 is selected by the wide area load distribution device 1-2. At this time, the wide area load distribution device 1-
No. 2 considers only the load of the Web server, and does not need to consider which P2P server has which content.

(3) The Web browser of the client 1-1 connects to the IP address returned by the wide area load distribution device 1-2 and performs an operation of acquiring contents.

(4) Web Server 1 Accessed
4 performs the same operation as a general Web server. That is, if the specified content exists, it is returned.
In the case of an external call such as GI or Servlet, it is activated and calls the P2P server 1-5.

(5) The P2P server 1-5 retrieves the specified content from the system. P2P server 1
In the case of -5, when the corresponding content exists by the search, the content is returned to the client 1-1. If it does not exist, return an error.

[Second Operation] Next, the P2P server 1-5
Has received a request from the Web server 1-4, the content of the request is not in the P2P server 1-5, and the P2P server 1
In the case where the content exists in -9, a procedure for acquiring the content will be described. FIG. 5 is a content acquisition sequence showing a flow of acquiring content.

(1) The content acquisition requester issues a content acquisition command with content information that can uniquely determine the content to be acquired as an argument. The content information to be specified does not need to be complete, as long as the content can be uniquely determined.

(2) The P2P server 1-5 having received the content acquisition command searches for the original content and the cache content managed by itself. As a result of this search, (a) if the original content exists, the content is returned to the caller. (B) If the cache content exists, return the content to the caller. (C) If neither the original content nor the cache content exists, the upper site (index proxy 1-6 or index server 1)
Send a search request to -7). In the case of the P2P server 1-5,
Since the higher-order site is the index server 1-7, a search request is issued to the index server 1-7.

Here, (C) is applicable when neither the original content nor the cache content exists.

(3) The search request is sent to the index server 1
If -7 is received, search the content information and return the result to the caller. Here, since the P2P server 1-9 holds the content satisfying the search request, the content information is stored in the index server 1-7. Therefore, the content information is returned from the index server 1-7 to the P2P server 1-5.

(4) The P2P server 1-5 that has received the content information acquires the content from the P2P server 1-9 described in the content information, and returns the content to the content acquisition instruction issue source. Then, a cache of the content is created, and a content information update notification is sent to the index server 1-7, which is a higher-level site. The index server 1-7 that has received the content information update notification registers the content information.

[Third Operation] Next, a procedure for updating content information will be described. Update of the content information is also performed on the P2P server. This P2P server may be the P2P server closest to the content information update commander. Here, the case where the original content exists in the P2P server 1-9 and the cache content exists in the P2P server 1-10 will be described. FIG. 6 is a content update sequence showing the flow of updating content.

(1) The content information update commander issues a content information update command using the content information to be updated as an argument.

(2) Upon receiving the content information update command, the P2P server 1-9 searches for content information that matches the content information targeted for the update command. If this content information is not uniquely determined, an error is returned.

Here, since the original content to be updated is managed by its own P2P server 1-9, the content information to be updated to a higher-level site (index proxy 1-6 or index server 1-7) is used as an argument. A content information update notification is performed, and the content information is updated. If the original content to be updated is managed by another P2P server, a content information update command is sent to the P2P server.

(3) Upon receiving the content information update notification, the index proxy 1-6 transmits to the higher-level site (the index proxy 1-6 or the index server 1).
-7, here, only the index server 1-7) is notified of the content update, and the content information is updated.

(4) Upon receiving the content information update notification, the index server 1-7 updates the content information. Then, the index server 1-7 is configured to execute all index proxies or P2P
Sends a content information update notification to the server.

(5) Upon receiving the content information update notification, the index proxy transmits a content information update notification to an index proxy or a P2P server that exists further below. Here, below the index proxy 1-6, the P2P servers 1-9, 1
Since only -10 exists, send to them.

(6) Upon receiving the content information update notification, the P2P server acquires (a) the original content from the original site when the corresponding cache content exists, and updates the cache. P2P
Since the server 1-10 has the cache content, the P2P server 1- which is the original site is
9 to obtain the original content and update the cache. Next, a content information update notification is performed, and the above processing is repeated. (A) When the corresponding cache content is not provided (here, the P2P server 1-9 and the P2P server 1-5), the content information update notification is ignored.

[Fourth Operation] Next, a procedure for adding a content will be described. Content is added to the P2P server. This P2P server may be the P2P server closest to the content registrant. here,
A case in which the P2P server 1-9 adds a content will be described. FIG. 7 is a content addition sequence showing the flow of adding content.

(1) The content registrant sends a content addition command using the content and its content information (storage directory, file name, etc.) as arguments to the P2P server 1.
Issue on -9.

(2) The P2P server 1-9 adds information such as a content size, a P2P server (own server) to be stored, and a creation date to the content information. Next, a content information update notification is performed using the content information added to the upper-level site (index proxy 1-6 or index server 1-7) as an argument, and the own database 2 is notified.
-4 and the hard disk 2-5 store the content information and the content.

(3) Upon receiving the content information update notification, the index proxy 1-6 stores the content information and notifies the upper-level site of the content information update notification.

(4) Upon receiving the content information update notification, the index server 1-7 stores the content information.

[Fifth Operation] Next, a procedure for deleting content will be described. The deletion of the original content and the cache content is performed on the P2P server. When content is deleted, all caches for that content in the system are also deleted. Further, when the cache content is deleted, only the cache is deleted, and other caches are not deleted. Here, P2
The original content exists on the P server 1-9, and the P2
A case will be described in which cache content exists in the P server 1-10 and the content deletion command is first transmitted to the P2P server 1-5. FIG. 8 is a content deletion sequence showing the flow when deleting content.

(1) The content deletion commander issues a content deletion command with content information having sufficient information to specify the content to be deleted as an argument. The content information to be specified does not need to be complete, and an instruction to delete the content by specifying a part of a character string (eg, a file name starting with XXX) or a range of a numerical sequence (eg, a file created after XXX date) is specified. It is possible to

(2) The P2P server 1-5 having received the content deletion command searches for a content that matches the content information of the argument. If the content is not uniquely determined, return an error. If the original content to be deleted is managed by its own P2P server 1-5, the upper site (index proxy or index server) is notified of the content deletion with the content information to be deleted as an argument, and the content is deleted.

Here, since the original content to be deleted is managed by the P2P server 1-9, a content deletion command is sent to the P2P server 1-9. The P2P server 1-9 having received this instruction searches for a content that matches the content information of the argument. P
Since the 2P server 1-9 manages the original content to be deleted by itself, the 2P server 1-9 notifies the upper site (the index proxy 1-6 or the index server 1-7) of the content deletion and deletes the content.

(3) The index proxy 1-6, which has received the content deletion notification, sends the content to the host site (the index proxy 1-6 or the index server 1-
7, here only the index server 1-7) is notified, and if a cache of the content information exists in its own database 2-4, the cache is deleted.

(4) Upon receiving the content deletion notification, the index server 1-7 deletes the content information. If the deleted content is the original, the index server 1-7 returns all the lower sites (index proxy or P2P server).
Sends an index change notification to

(5) Upon receiving the index update notification, the index proxy deletes the cache of the content information if the cache of the content information exists. Then, an index change notification is transmitted to the lower site (index proxy or P2P server).

(6) When the P2P server receives the index change notification, if the cache content of the corresponding content information exists, the P2P server deletes the content using the content information of the cache content as an argument at the upper-level site (index proxy or index server). Notify and delete cache contents.
Here, since the P2P server 1-10 has the cache content, if the P2P server 1-10 receives the index change notification, the P2P server 1-10 performs a content deletion notification using the content information of the cache content as an argument to the upper-level site, Delete cache contents.

(7) Upon receiving the index update notification, the index server 1-7 deletes the corresponding content information cache and transmits an index change notification to the lower site.

[Sixth Operation] When the P2P server is newly started, the fact that the P2P server has been started is notified to another P2P server.
There is no need to notify the server or notify the index server 1-7. This is because the P2P server immediately after the start is started without any original content or cache content. When the P2P server is restarted after being down once, it automatically re-registers the contents that it had before.

When a new site is added, the following processing is performed.

1. Top site (index proxy or index server) on P2P server in site
Set.

2. Register the P2P server in the lookup server 1-8.

3. Start the Web server.

With the above processing, a new site can be easily added. Since the index server, the index proxy, and the P2P server are very loosely coupled as described above, it is very easy to change the configuration of the system.

In the case of deleting a site, it is only necessary to issue a deletion notification to the higher-level site. Lookup server 1-8
Searches independently to determine whether the site is currently running.

An index proxy can also be dynamically added / deleted. Add index proxy /
Deletion is possible only by changing the upper site of the tree structure of the system. There is no need to change information such as index information held by the index server and contents held by the P2P server.

As described above, the index proxy, P
A 2P server or the like can be easily added / deleted as needed.

[Effects of the Embodiment] As described above, the cache content for the original can be easily copied to another P2P server as needed, and the original content and the cache content can be easily deleted. In addition, an index proxy, a P2P server, and the like can be easily added or deleted at an arbitrary position in the system as needed. As a result, the target content can be easily and reliably accessed from any site into the system.

Since the desired contents can be easily and reliably accessed, even if a part of the P2P server goes down or a part of the network goes down, the whole system does not go down and continues. Service can be provided.

Since the down P2P server is automatically deleted from the search target, only the information of the currently activated P2P server can be searched, and efficient access is possible.

By adding an index proxy or the like as needed to assist the index server, even if traffic cannot be connected to the index server in a concentrated manner, the cache of index information of the P2P server and the index proxy can be used. By using, it is possible to easily and reliably access the target content.

When traffic is concentrated on a part of the site, the system is stopped by dynamically adding a P2P server to the site or creating a cache of the content where the traffic is concentrated on another site. It is possible to distribute the load without any problem.

As the system is operated, the load on the index server may increase unexpectedly. If the processing of the index server becomes a bottleneck, the system is stopped by dynamically adding an index proxy. It is possible to reduce the load concentration on the index server without performing.

[0088]

As described in detail above, according to the content distribution management system of the present invention, it is possible to flexibly respond to various situations such as traffic concentration. As a result, the target content can be easily and reliably accessed, and the communication quality can be improved.

[Brief description of the drawings]

FIG. 1 is an overall configuration diagram showing a content distribution management system according to the present invention.

FIG. 2 is a configuration diagram illustrating a P2P server.

FIG. 3 is a configuration diagram showing an index proxy and an index server.

FIG. 4 is an access sequence showing a flow of acquiring content.

FIG. 5 is a content acquisition sequence showing a flow of acquiring content.

FIG. 6 is a content update sequence showing a flow of updating content.

FIG. 7 is a content addition sequence showing a flow of adding content.

FIG. 8 is a content deletion sequence showing a flow when deleting content.

[Explanation of symbols]

 1-1: Client 1-2: Wide area load distribution device 1-3: Site 1-4: Web server 1-5, 1-9, 1-10: P2P server 1-6: Index proxy 1-7: Index server 1-8: Lookup server 2-1, 3-1: Distributed environment 2-2, 3-2: P2P module 2-3, 3-3: Admin module 2-4, 3-4: Database 2-5: hard disk

Continued on the front page F term (reference) 5B045 BB42 DD17 GG04 5B075 KK03 KK07 KK13 KK34 KK39 PQ02 PQ42 5B082 HA02 5B089 GA11 GA11 JA19 JA33 KA06 KB04 KC44 MA03

Claims (3)

[Claims] 1. A content distribution management system for providing uniform content on a network distributed over a wide area, a wide area load distribution device for selecting a lower site in consideration of a lower site, a load on a network, and the like; And a content distribution server interconnected by peer-to-peer, and an index server that exists at the top and manages all content information in the system. A content distribution management system comprising: 2. The content distribution management system according to claim 1, further comprising an index proxy that caches the content information and supplements the index server, thereby avoiding load concentration on the index server. Characteristic content distribution management system. 3. The content distribution management system according to claim 1, wherein each server in the system is determined whether or not each server is currently running, and if not, the server is deleted from the list and a new server is registered. A content delivery management system comprising a look-up server that adds the information to a list and notifies the index server.