JP2005018293A - Content delivery control device, content delivery control method and content delivery control program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To realize a content delivery system that reduces loads on servers and networks. <P>SOLUTION: A content acquisition request reception part 310 receives content acquisition requests from users. A classification control part 324 classifies the content acquisition requests according to information about cache servers to be assigned and the like. A batching time computation part 326 sets batching times according to the classification. For every piece of content by the content acquisition requests, multicast streams are generated at periods of the batching time, and the users receive desired content. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a content distribution technique on a network.
[0002]
[Prior art]
In recent years, content distribution services using multimedia data such as voice and moving images as well as text data such as WWW (World Wide Web) and e-mail have been provided along with the broadbandization of the Internet. Among them, video-on-demand (hereinafter referred to as “VOD”) is home entertainment, digital video library, movie on demand, distance learning, remote conference, Internet shopping, news on demand, It is expected to be used in various fields such as medical information services.
[0003]
[Non-Patent Document 1]
Yasushi Ikeda and 4 others, “Configuration and Verification of Policy Selective CDN Platform”, IEICE Transactions Society, The Institute of Telecommunications Information, March 2003, VOL. J86-BNo. 3, p. 400-409
[Non-Patent Document 2]
Huadong Ma, Kang G., et al. Shin, “Multicast Video-on-Demand Services”, ACM SIGCOMM Computer Communication Review Volume 32, Issue 1, pp. 199 31-43, January 2002
[0004]
[Problems to be solved by the invention]
Multimedia data including images and the like has not only a large amount of data compared to text data, but also has a high transmission cost because it is necessary to ensure a continuously stable network bandwidth. Accordingly, there are many distribution requests from users, and if the network and server processing capabilities are exceeded, service delays and stops occur.
[0005]
As a technique proposed to deal with such a problem, there is a CDN (Content Delivery / Distribution Network). CDN arranges a plurality of cache servers on a network for content distribution. These cache servers hold a copy of the content. The CDN is configured to check the URL (Uniform Resource Locator) of the client terminal accessed by the user and connect to the nearest cache server. Since content distribution to users is performed by a plurality of cache servers, the load on the network and the cache server is distributed.
[0006]
However, even if the CDN is used, if the number of users, the number of contents, and the data amount of the contents themselves increase, the cache server must be increased each time. CDN is not a decisive action for a content distribution system to break away from a high cost structure. Further, since CDN is content distribution based on unicast communication, an increase in the number of users directly leads to an increase in the number of communication connections. Since the load on the network becomes large, when the number of accesses to the cache server increases, the allowable range of the network is often exceeded.
[0007]
The present invention has been made in view of the above background, and an object of the present invention is to provide a technique for distributing content while suppressing a load on a network or a cache server.
[0008]
[Means for Solving the Problems]
One embodiment of the present invention relates to a content distribution control apparatus. This apparatus receives from a client terminal connected via a network an acquisition request receiving unit that receives a content acquisition request that is an acquisition request for desired digital content configured by including image or audio data. For each content acquisition request, a cache server selection unit that allocates a cache server to which the requested digital content is to be transmitted, and batching that sets a predetermined waiting time for each digital content to be transmitted by the allocated cache server A time setting unit and a cache server control unit that transmits digital content from the allocated cache server to the client terminal when the set batching time has elapsed are included.
[0009]
“Cache server” refers to a server device having a function of transmitting digital content to a client terminal. A plurality of cache servers may have the ability to distribute the same digital content. Moreover, the cache server may have a capability of distributing a plurality of digital contents by one unit.
The content distribution control device is a device for connecting a client terminal and a cache server to control the entire content distribution system.
[0010]
The cache server selection unit of this device may assign a cache server based on the digital content requested in the received content acquisition request.
[0011]
For example, in the case of digital contents for which demand from a large number of users is expected, there is a possibility that only one cache server cannot meet the demands of all users. In such a case, the load on the cache server or the network can be distributed by distributing content acquisition requests for the same digital content to a plurality of cache servers.
[0012]
The cache server selection unit of this apparatus may assign a cache server based on the received information regarding the client terminal that has made the content acquisition request.
[0013]
“Information about the client terminal” means the IP address of the client terminal, the domain name to which the client terminal belongs, device information about the performance of the client terminal, the contents of the contract for the content distribution service, information about the network to which the client terminal belongs. Hereinafter, the information regarding the client terminal is simply referred to as “client information”.
[0014]
For the digital content requested in the received content acquisition request, when the content acquisition request has already been made, the client terminal that has made the received content acquisition request is assigned to the content acquisition request that has already been made. If the network distance to the cache server is equal to or greater than a predetermined value, the cache server selection unit may assign a cache server different from the assigned cache server to the received content acquisition request.
[0015]
“Network distance” refers to a communication distance between two communication nodes. The network distance may be defined by the number of routers and networks that pass through. Alternatively, it may be defined by a physical distance between the cache server and the client terminal, or may be defined in consideration of communication throughput.
[0016]
It further includes an affiliated network determination unit that detects whether or not the client terminal that has received the received content acquisition request belongs to a network that supports multicast, and the affiliated network determination unit includes the client terminal that belongs to the multicast network. If it is detected that the data is not stored, the cache server control unit may transmit the digital content from the cache server to the client terminal even if the batching time has not elapsed.
[0017]
The server further includes a server load detection unit that detects server load information that is information related to the processing load applied to the cache server, and the cache server selection unit is a digital content requested in the received content acquisition request based on the detected server load information. May be assigned to the cache server to be transmitted.
[0018]
“Server load information” refers to the usage rate of the CPU (Central Processing Unit) of the cache server, the memory usage and memory occupation rate of the recording medium such as RAM (Random Access Memory), and the number of network connections processed by the cache server. It may be. Further, the processing load may be determined in consideration of the processing capacity of the cache server. In addition to this server load information, information related to the cache server, including the server IP address, the domain name to which the server belongs, device information related to the performance of the cache server, and the contents of content having distribution capability, etc. This is called “server information”.
When the load is concentrated on a specific cache server, the load on the cache server can be distributed as a whole by sharing the processing with another cache server.
[0019]
The batching time setting unit of this device may change the length of the batching time based on the server load information detected for the allocated cache server.
[0020]
The batching time setting unit of this apparatus may set the batching time based on the number of content acquisition requests for digital content to be transmitted by the allocated cache server.
[0021]
The cache server selection unit further includes a first policy setting unit that detects an input of a first policy that is a rule on how to allocate a cache server by the cache server selection unit, and based on the first policy that has detected the input, the cache server selection unit A cache server may be allocated.
[0022]
The batching time setting unit further includes a second policy setting unit that detects an input of the second policy, which is a rule for setting the batching time by the batching time setting unit, and the batching time setting unit is based on the second policy in which the input is detected. You may set batching time.
[0023]
Another aspect of the present invention is also a content distribution control device. This apparatus receives from a client terminal connected via a network an acquisition request receiving unit that receives a content acquisition request that is an acquisition request for desired digital content configured by including image or audio data. A content detection unit that detects digital content requested in the content acquisition request, an acquisition request classification unit that classifies the received content acquisition request for each detected digital content, and temporarily holds the classified content acquisition request The acquisition request holding area that is a memory area of the acquisition request holding area management unit that reserves in the recording medium for each classification and the content acquisition request group that holds each acquisition request holding area in the acquired acquisition request holding area. Digital content requested by the request For each content acquisition request group held in each of the acquired acquisition request holding areas and a cache server setting unit that allocates a cache server to be transmitted, a batching time that is a time for temporarily holding those acquisition requests is set. The batching time setting unit to monitor, the batching time management unit to monitor the progress of the set batching time, and when the set batching time has passed, it is held in the acquisition request holding area corresponding to the batching time. A cache server control unit that reads out the content acquisition request that is received and distributes the digital content to the client terminal that has made the content acquisition request read from the cache server assigned in correspondence with the acquisition request holding area.
[0024]
Yet another aspect of the present invention is a content distribution control method. The method includes a step of receiving a content acquisition request, which is an acquisition request for desired digital content configured by including image or audio data, from a client terminal connected via a network, and the received content acquisition request Assigning a cache server to which the requested digital content is to be transmitted, setting a batching time which is a predetermined waiting time for each digital content to be transmitted by the allocated cache server, and setting the batching time. It is characterized by including a step of transmitting digital content from the cache server to the client terminal when triggered.
[0025]
It should be noted that any combination of the above-described constituent elements and a representation of the present invention converted between a method, apparatus, server, system, computer program, recording medium, etc. are also effective as an aspect of the present invention.
[0026]
DETAILED DESCRIPTION OF THE INVENTION
On the Internet, users often rush to specific servers. In particular, a large amount of access tends to be temporarily concentrated on a server having popular digital content (hereinafter simply referred to as “content”). If accesses from users concentrate on a specific server, the load on the server increases and the responsiveness deteriorates. Moreover, the rapid increase in communication traffic tends to cause local congestion in the network. In recent years, broadbandization has enabled so-called streaming in which content with a large amount of data such as voice and moving images is received and played back simultaneously. Even if the network becomes broadband, the load on the server and network increases incidentally.
[0027]
There is IP multicast (hereinafter simply referred to as “multicast”) as a technique for reducing transmission costs on an IP network. Multicast is a data distribution technique in which a plurality of clients can receive the same data with a single data transmission. Client terminals that are the targets of multicast distribution are registered in the multicast group. If the multicast distribution packet received from the server is a packet to be transmitted to each client terminal belonging to a different network, the router copies the packet and sends it to each network. In unicast, the server establishes a communication connection to each client to receive a service, and transmits the same packet to each client. On the other hand, in multicast, since a router copies a packet sent by a server as necessary, communication traffic as a whole network is reduced. In addition, since the server does not require processing for establishing a communication connection for each client, the load on the server itself is reduced.
[0028]
The advantages of multicast are summarized as follows.
(1) Since there is no need to establish a communication connection for each client, the communication bandwidth consumption of the network is reduced. For this reason, the communication traffic of the whole network is also reduced.
(2) Since there is no need to establish a communication connection for each client, the load on the server is reduced. For this reason, the throughput of the entire content distribution system is improved.
(3) Since it is possible to cope with a new client only by registering the new client in the multicast group, the entire content distribution system is highly scalable.
[0029]
However, in the existing multicast distribution method, since a plurality of clients participate in the same multicast stream, it is necessary to access the server at the same time. Therefore, multicast is not suitable for applications such as VOD systems in which users access at various times, and multicast / VOD at the IP level has not become widespread.
[0030]
The current situation surrounding VOD is not necessarily rock-solid, and a new VOD architecture is required. The content distribution control apparatus proposed in the present invention pays attention to multicast at the IP level, and proposes a new multicast VOD (hereinafter referred to as “MVOD: Multicast VOD”) architecture.
[0031]
There is a method called a batching method as a method for making multicast distribution close to VOD. In the present embodiment, MVOD is realized by incorporating this batching method into the content distribution control device. The content distribution control device is a device for controlling a content distribution system that distributes a content acquisition request from a user to an appropriate cache server.
[0032]
FIG. 1 is a conceptual diagram of a content distribution system 200. The content distribution system 200 includes a cache server group 210, a client terminal group 220, and a content distribution control device 230. The cache server group 210 is a set of cache servers that perform content distribution, and the client terminal group 220 is a set of client terminals that should receive the content distribution service from the cache server. The content distribution control device 230 is a device that mediates both. The content distribution control device 230 searches the cache server group 210 for a cache server that holds the content requested by the client terminal. In addition, among the searched cache servers, a cache server to which content is actually distributed to the client terminal is also selected. Further, the selected cache server also controls the timing of distributing the requested content to the client terminal.
[0033]
Hereinafter, the batching method will be described first, and the advantages of incorporating the batching method into the content distribution control device will be described. Thereafter, processing of the entire content distribution system will be described.
[0034]
The “batching method” is a method of processing content acquisition requests from a plurality of clients for the same content with a single multicast stream for a certain period of time. Hereinafter, this fixed time is referred to as “batching time”. Recruiting content acquisition requests during batching time is called “batching”. A server for distributing content is provided with a queue for temporarily storing content acquisition requests from clients. A multicast group is generated for each requested content for a plurality of clients that are transmission sources of content acquisition requests stored in the queue. Then, a multicast stream is periodically generated with the set batching time as a cycle, and content distribution to the client terminal is realized.
[0035]
In the multicast distribution by the batching method, even if there are many content acquisition requests during the batching time, it can be handled by one multicast stream, so the scalability against the increase in the number of clients is high. The server load is also reduced compared to unicast. The user can receive the content by any multicast stream at the timing when the user accesses the content distribution server. The longer the batching time, the higher the throughput of the content distribution system. This is because the frequency at which the server generates a multicast stream and distributes the content decreases, thereby reducing the load on the server. For the same reason, communication traffic in the network is also reduced. If the batching time is set short, the immediacy of the user's content acquisition request is improved. FIG. 2 shows a conceptual diagram of the batching method.
[0036]
According to the batching method, content distribution can be performed while taking advantage of multicast. However, since the batching method is premised on implementation on each server for content distribution, it is necessary to add a function for batching to each server, and the equipment cost is high. In addition, each of the content distribution servers needs to perform various processes such as queue management and batching time management for user content acquisition requests. This increases the processing load on each server.
[0037]
In the batching method, a queue management method for storing user requests is also an important factor. Typical queue management methods include FCFS (First Come First Server), MQL (Maximum Queue Length first), and MFQ (Maximum Factored Queue First first). However, these methods are not efficient and are not suitable for environments with a large number of contents. An existing queue management method is suitable for managing a queue in a content distribution server, but a device that distributes content acquisition requests from client terminals to a plurality of cache servers, such as the content distribution control device in the present embodiment. It is because it does not fit. This is because the performance of the content distribution control device is affected by factors such as the load on each cache server and network. These methods are not queue management methods assuming such a situation. Therefore, it is necessary for the content distribution control apparatus to adopt a queue management method that takes these factors into consideration. The queue management of the content distribution control device will be described in detail later.
[0038]
In the present embodiment, a content distribution control device incorporating a batching method and a content distribution system using the content distribution control device are proposed. A queue for temporarily storing content acquisition requests from clients is prepared in the content distribution control device, and the cache server is controlled based on the queue management method proposed next. The advantages of adopting the batching method in the content distribution control device are summarized into six, and are listed below.
[0039]
(1) It can be realized at low cost.
No complicated processing related to batching is required in the cache server. Therefore, it is not necessary to add functions to the cache server, and the existing cache server can be used as it is. When batching is performed on the cache server, the queue must be held on the cache server. Therefore, the capacity of each queue must be designed to peak in all the cache servers. However, when batching is performed in the content distribution control apparatus, peak design may be performed only in the content distribution control apparatus. As described above, this advantage is significant because the cost issue is very important in content distribution.
[0040]
(2) Queue management can be performed flexibly.
Even when the content distribution system is in operation, the queue settings can be changed dynamically according to the cache server and network conditions. That is, the queue can be controlled in consideration of various situations such as the load on the cache server and the network, the number of content acquisition requests for content, and the distribution status of client terminals that are making content acquisition requests. In addition, since all content acquisition requests are centrally processed by the content distribution control device, comprehensive queue management is possible with a plurality of cache servers instead of a single cache server. Since settings related to the queue management policy can be made centrally, the maintainability of the entire system is improved.
[0041]
(3) Control according to the status of the client terminal is possible.
The client information is acquired by the content distribution control device, and the cache server can be allocated based on the client information. For example, the quality of the distribution service can be improved by allocating the cache server so that the client server receiving the service and the client terminal receiving the service are close to each other to some extent on the network. Alternatively, if the content has regional characteristics such as local news, the distribution status of client terminals that require the content can be predicted in advance. In such a case, efficient content distribution is possible by allocating a cache server based on the content content and the location of the client terminal on the network.
[0042]
(4) Address management becomes easy.
The content distribution control device can centrally grasp and manage the multicast addresses being used in a plurality of cache servers. In addition, in an environment where a multicast stream can be allocated only in the content distribution control device, it is easy to prevent the use of an illegal multicast address in the content distribution system.
[0043]
(5) High scalability.
As the number of cache servers increases, only the setting of the content distribution control device needs to be changed, so the scalability of the scale is high. Further, since the content distribution control apparatus only assigns the address or URL of the cache server and does not need to perform the content distribution process by itself, it can be operated independently of the multicast protocol.
[0044]
(6) Partial applicability is high.
Based on the client information, the content distribution control device can determine whether the network to which the client terminal belongs is a multicast network or a non-multicast network. The allocation of the cache server can be controlled by this discrimination information. Since the content distribution control device controls the allocation of the cache server, even if a multicast server that does not support multicast is mixed in the content distribution system, it can be managed collectively.
[0045]
FIG. 3 is a diagram showing an outline of the processing flow of the content distribution control apparatus. In the figure, some client terminals belong to a network that supports multicast and some do not.
(1) The client terminal makes a content acquisition request (Query of Contents) to the content distribution control device (CRS with Batching). At this time, the client terminal obtains the location of the content distribution control device by accessing a DNS (Domain Name System) server in advance.
(2) The content distribution control device recruits content acquisition requests during the batching time, that is, performs batching, and generates a multicast group. If the client terminal does not belong to a network that supports multicast, a unicast cache server is allocated as a non-batch target.
(3) When the batching time has elapsed after the content distribution control device solicits the content acquisition request, a redirect instruction to the allocated cache server is issued to the client terminal that has made the content acquisition request accepted so far. . The redirect command notifies the client terminal of the location of the cache server to which the content is to be distributed.
(4) The client terminal requests distribution of content to the cache server designated by the redirect command. Hereinafter, in order to explicitly distinguish the content acquisition request from the client terminal to the content distribution apparatus and the content distribution request from the client terminal to the cache server, the former is particularly referred to as the “first request” and the latter according to the situation. Also referred to as “second request”.
(5) The cache server delivers the requested content to the requesting client terminal by multicast or unicast.
As described above, MVOD for a plurality of client terminals is realized by the processes (1) to (5).
[0046]
It is assumed that a content acquisition request is made beyond the processing capacity of the cache server due to an increase in content or an increase in capacity. In this case, the number of client terminals that can be handled simultaneously in the entire VOD system varies depending on which content acquisition request is assigned to which cache server in which order. The queue management method is an important element in batching because it is directly linked to allocation scheduling.
[0047]
Typical queue management methods include FCFS, MQL, and MFQ, which are premised on queue management on a server that performs content distribution. Therefore, when these queue management methods are applied as they are in the content distribution control device, content acquisition requests from a plurality of clients are allocated to a plurality of cache servers without considering the load on each cache server or network. Therefore, efficiency is not good. In view of this, in the content distribution apparatus according to the present embodiment, a queue management method is proposed in consideration of the network wideness connecting the cache server and the client terminal and the number of contents.
[0048]
First, the various queue management methods described above will be described.
(1) FCFS
The content acquisition request that has been waiting for the longest time in the queue is set as the next allocation target. The advantage of this method is that it is fair to the content acquisition requests of all client terminals because the popularity of the content is not considered. However, this method reduces the throughput of the VOD system.
(2) MQL
The content with the highest content acquisition request in the queue is set as the next allocation target. This method does not take into consideration the time when the content acquisition request is made and the status of the client terminal, and uses only the number of content acquisition requests in the queue as a parameter, so there is no fairness. Content distribution is performed preferentially for more popular content.
(3) MFQ
The next allocation target is determined by the access frequency considering both the time waited in the queue and the popularity of the content. This method can realize a throughput close to MQL while maintaining a certain degree of fairness.
[0049]
So far, three existing queue management methods have been described, but these are inefficient in a system having a plurality of cache servers, and are inappropriate when batching is performed in a content distribution control apparatus. This is because these queue management methods are assumed to be implemented on a single server for content distribution, and thus the wide area of the network connecting the cache server and the client terminal is not considered. Specifically, there are three problems. The summary is listed below.
(1) Since all content acquisition requests are managed by one queue, fine queue management cannot be performed for each content based on the load on the cache server and the number of content acquisition requests.
(2) Since the queue management parameter is only the number of content acquisition requests and the waiting time until the processing of the requests, the location information of the cache server and the client terminal cannot be taken into consideration. Information on the network distance to the client terminal cannot be reflected.
(3) Since the batching cycle is constant and all the contents have the same batching time, the batching efficiency for one content cannot be considered.
In consideration of the above problems, FIG. 4 shows an outline of a new queue management method proposed in the present embodiment, which corresponds to the batching process portion of FIG.
[0050]
FIG. 4 is a schematic diagram for explaining the flow of the batching process. The batching processing unit accumulates content acquisition requests from users and assigns a cache server to which desired content is distributed to each user. Here, before explaining the functional block diagram of the batching processing unit, an outline of the processing will be described.
[0051]
The first request (Query) transmitted from the client terminal to the content distribution control apparatus first enters a common queue. Next, a classification unit (Classifier) generates a queue for each content based on status information (client information, server information, content information), assigns a corresponding cache server, and sets a batching time for each. “Content information” refers to content popularity and information attached to the content. When the batching time becomes zero, the redirection processing unit (Redirect Engine) issues a redirect command to each client terminal by setting the client terminals that are the transmission source of the first request in the queue as one multicast group. Processing from when the redirect command is transmitted until the allocated cache server performs content distribution will be described in detail later.
[0052]
The flow of processing in the classification unit is shown in FIG.
The received first request is taken out by the classification unit. And it is judged whether the client terminal which transmitted the taken out 1st request belongs to the network corresponding to a multicast. If multicast is not supported, the content is controlled to be unicasted by an appropriate cache server. If the client terminal belongs to a multicast-compatible network, the cache server allocated for distributing the content requested by the first request and the client terminal that is the transmission source of the newly received first request Check network distance. When the network distance is larger than a predetermined value, a new queue is generated, and a cache server having a reasonable proximity is allocated for content distribution. Then, the batching time is set based on the server information and content information of the allocated cache server and policy information described later.
[0053]
On the other hand, when the network distance between the already allocated cache server and the client terminal that is the transmission source of the newly received first request is smaller than a predetermined value, the load status of the cache server is checked. If the load on the cache server is equal to or less than a predetermined value, the newly received first request is added to the queue to which the cache server is assigned as it is.
[0054]
The batching time set when the classification unit generates a queue is mainly determined by the load status of the cache server and the popularity of the content. When the load on the cache server tends to increase, the number of multicast streams processed by the cache server is reduced by increasing the batching time to suppress the increase in load. Thereby, since the number of users processed with one multicast stream increases, the efficiency of batching becomes high. On the other hand, when the load on the cache server is light, the immediacy of content delivery with respect to the request is improved by shortening the batching time.
[0055]
The content distribution control apparatus can assign a cache server using not only the position of the cache server and the client terminal on the network but also the load of the cache server and the content name as parameters. The configuration of the content distribution control device is shown in FIG.
[0056]
There are four components of the content distribution control device: SNMP (Simple Network Management Protocol), policy, mapping processing unit, and batching processing unit. These will be described in order.
(1) SNMP
SNMP is composed of three components, an agent, a manager, and a management block (MIB: Management Information Base), and is a protocol that manages and grasps the status of network devices scattered on the network. In the present embodiment, it is used for grasping the state of each cache server. The SNMP manager accesses a cache server group (Neighboring Servers), acquires each server information, and stores it in the server information storage unit.
(2) Policy
The policy describes how the content distribution control apparatus should perform cache server allocation. Specifically, it is a function that performs cache server selection using server information, client information, and content information as parameters. A plurality of policies can be mounted on the content distribution control apparatus and can be dynamically changed remotely by an administrator.
(3) Mapping processing unit
The mapping processing unit has two functions. One is a function of processing status information acquired by SNMP or the like based on a policy and selecting a cache server to be allocated. Status information refers to client information, content information, and server information stored in a client information storage unit, a content information storage unit, and a server information storage unit, respectively. The other is to provide status information required when the classification unit performs queuing in the batching process.
(4) Batching processing unit
The batching processing unit serves as an interface with the client terminal. The batching processing unit receives a content acquisition request from the client terminal, and transmits a redirect command after batching. The batching processing unit is an actual implementation of the queue management method described above.
[0057]
FIG. 7 is a sequence diagram until the client terminal makes the first request and receives the multicast stream.
First, the client makes a first request to the content distribution control device using HTTP (HyperText Transfer Protocol) (GET: Content). The content distribution control apparatus performs batching and issues a metafile redirect command to the client terminal (Redirect: Metafile). A metafile is a file that describes information necessary for the browser of the client terminal to perform streaming playback. The metafile includes a URL that is the location information of the allocated cache server and content. In addition, information regarding the content abstract, title, and copyright may be included.
[0058]
In accordance with the redirect command, the client terminal requests and acquires a metafile from the content distribution control device (GET: Metafile, RES: Metafile). At this time, all client terminals batched into the same multicast stream acquire the same metafile. The client terminal makes a second request to the allocated cache server based on the contents described in the metafile (GET: Content). When the cache server receives the request, it starts multicast stream distribution.
[0059]
When the content is distributed by unicast, the content distribution control device causes the client terminal to download a unicast metafile. Then, the client terminal makes a second request to the designated cache server by unicast.
[0060]
FIG. 8 is a functional block diagram of the batching processing unit 232. The batching processing unit 232 includes a content acquisition request receiving unit 310, a classification processing unit 320, and a redirection processing unit 330. The content acquisition request receiving unit 310 receives a content acquisition request from the client terminal and temporarily holds it. The classification processing unit 320 classifies the content acquisition request group received by the content acquisition request receiving unit 310 based on the status information. The classification processing unit 320 includes a classification control unit 324 and a batching time calculation unit 326. The classification control unit 324 appropriately takes out and classifies content acquisition requests received and held by the content acquisition request receiving unit 310. The content acquisition request includes information such as the IP address of the client terminal that is the transmission source and the requested content. Based on these pieces of information, the mapping processing unit 350 selects an optimal cache server for delivering desired content to the client terminal that is the transmission source. The batching time calculation unit 326 calculates an optimum batching time based on the status information received from the classification control unit 324.
[0061]
The redirection processing unit 330 controls content distribution to the client terminal by the cache server specified by the classification processing unit 320. The redirection processing unit 330 includes a content acquisition request storage unit 332, a content acquisition request processing unit 334, a batching time processing unit 336, a metafile generation unit 338, a redirect command transmission unit 340, a metafile transmission unit 342, and a metafile storage unit 344. Including. The content acquisition request processing unit 334 acquires the content acquisition request classified by the classification control unit 324. The content acquisition request processing unit 334 stores the content acquisition request in the content acquisition request storage unit 332 according to the classification. In response to an instruction from the content acquisition request processing unit 334, the batching time processing unit 336 sets the batching time as a timer and manages the time. The metafile generation unit 338 generates a metafile and stores it in the metafile storage unit 344.
[0062]
When the batching time processing unit 336 detects that the batching time has elapsed, the redirect command transmission unit 340 issues a redirect command to the client terminal. When the metafile transmission unit 342 receives a metafile acquisition request from the client terminal, the metafile transmission unit 342 transmits the metafile stored in the metafile storage unit 344. As described above, the client terminal makes a second request to the cache server according to the description of the received metafile, and the content is distributed.
[0063]
FIG. 9 is a functional block diagram of the mapping processing unit 350. The mapping processing unit 350 includes a status information acquisition unit 352, a multicast network determination unit 354, and a cache server selection unit 356. The status information acquisition unit 352 acquires status information and information related to the set policy. The multicast network determination unit 354 determines whether the client terminal that has made the first request request belongs to a multicast-compatible network. In response to the first request, the cache server selection unit 356 selects a cache server on which content distribution is to be performed.
[0064]
As described above, in the embodiment, it is only necessary to add the content distribution control device to the existing content distribution system. In addition, mounting is easy, low cost, and excellent scalability. For this reason, it is considered that large-scale VOD distribution is possible.
[0065]
The present invention has been described based on the embodiments. The embodiments are exemplifications, and it will be understood by those skilled in the art that various modifications can be made to combinations of the respective constituent elements and processing processes, and such modifications are within the scope of the present invention. .
[0066]
As such a modification, the cache server may not generate a multicast stream with a batching time as a period, but may be generated when a predetermined number of first requests for a certain content are received. The batching time may be changed according to the time zone for multicast distribution. For example, the batching time may be set longer in a time zone where the network load during the day is generally high, and the batching time may be set shorter in a time zone where the network load is generally low, such as midnight.
[0067]
Further, the policy regarding the length of the batching time and the cache server allocation method may be set in a program format by combining a plurality of functions.
[0068]
【The invention's effect】
According to the present invention, it is possible to distribute content while suppressing a load on a network or a cache server.
[Brief description of the drawings]
FIG. 1 is a schematic diagram showing the role of a content distribution control apparatus in the present embodiment.
FIG. 2 is a schematic diagram for explaining a batching method.
FIG. 3 is a diagram showing an overview from allocation of a cache server to content distribution by the content distribution control device;
FIG. 4 is a diagram showing an outline of a flow of batching processing.
FIG. 5 is a flowchart of processing in the classification unit of FIG. 4;
FIG. 6 is a diagram illustrating a configuration of a content distribution control device.
FIG. 7 is a sequence diagram until a client terminal makes a content acquisition request and receives a multicast stream.
FIG. 8 is a functional block diagram of a batching processing unit in the content distribution control apparatus.
FIG. 9 is a functional block diagram of a mapping processing unit in the content distribution control device.
[Explanation of symbols]
230 content distribution control device, 232 batching processing unit, 310 content acquisition request receiving unit, 320 classification processing unit, 324 classification control unit, 326 batching time calculation unit, 330 redirection processing unit, 332 content acquisition request storage unit, 334 content acquisition request Processing unit, 336 Batching time processing unit, 340 Redirect command transmission unit, 342 Metafile transmission unit, 350 mapping processing unit, 352 Status information acquisition unit, 354 Multicast network determination unit, 356 Cache server selection unit.

Claims (14)

An acquisition request receiving unit that receives a content acquisition request, which is an acquisition request for desired digital content configured by including image or audio data, from a client terminal connected via a network;
For each received content acquisition request, a cache server selection unit for assigning a cache server to which the requested digital content is to be transmitted;
A batching time setting unit for setting a batching time which is a predetermined waiting time for each digital content to be transmitted by the allocated cache server;
And a cache server control unit configured to transmit the digital content from the allocated cache server to the client terminal when the set batching time has elapsed. The content delivery control apparatus according to claim 1, wherein the cache server selection unit assigns the cache server based on digital content requested in the received content acquisition request. The content delivery control apparatus according to claim 1, wherein the cache server selection unit assigns the cache server based on information about the client terminal that has made the received content acquisition request. For the digital content requested in the received content acquisition request, when a content acquisition request has already been made, the client terminal that has made the received content acquisition request is allocated to the content acquisition request that has already been made If the network distance to the assigned cache server is equal to or greater than a predetermined value, the cache server selection unit selects a cache server different from the assigned cache server for the received content acquisition request. 4. The content distribution control device according to claim 1, wherein the content distribution control device is assigned. An affiliated network determination unit that detects whether or not the client terminal that has made the received content acquisition request belongs to a network that supports multicast;
If the affiliated network determination unit detects that the client terminal does not belong to the multicast network, the cache server control unit determines whether the batching time has elapsed or not from the cache server to the client terminal. 5. The content distribution control device according to claim 1, wherein the digital content is transmitted to the device. A server load detection unit that detects server load information that is information relating to the processing load applied to the cache server;
6. The cache server selection unit according to claim 1, wherein the cache server selection unit assigns a cache server to which digital content requested in the received content acquisition request is to be transmitted based on the detected server load information. The content distribution control device described in 1. The content delivery control apparatus according to claim 6, wherein the batching time setting unit changes the length of the batching time based on server load information detected for the allocated cache server. The said batching time setting part sets the said batching time based on the number of the said content acquisition requests with respect to the digital content which the said allocated cache server should transmit. Content delivery control device. A first policy setting unit for detecting an input of a first policy that is a rule on how to allocate the cache server by the cache server selection unit;
The content distribution control apparatus according to claim 1, wherein the cache server selection unit performs the allocation of the cache server based on a first policy in which the input is detected. A second policy setting unit for detecting an input of a second policy that is a rule for setting the batching time by the batching time setting unit;
The content delivery control apparatus according to any one of claims 1 to 9, wherein the batching time setting unit sets the batching time based on a second policy in which the input is detected. An acquisition request receiving unit that receives a content acquisition request, which is an acquisition request for desired digital content configured by including image or audio data, from a client terminal connected via a network;
A content detection unit for detecting digital content requested in the received content acquisition request;
An acquisition request classifying unit that classifies the received content acquisition request for each detected digital content;
An acquisition request holding area management unit for securing an acquisition request holding area, which is a memory area for temporarily holding the classified content acquisition requests, in a recording medium for each classification;
For each content acquisition request group held in each of the secured acquisition request holding areas, a cache server setting unit that allocates a cache server to which the digital content requested by the content acquisition request is to be transmitted;
For each content acquisition request group held in each of the secured acquisition request holding areas, a batching time setting unit that sets a batching time that is a time for temporarily holding those content acquisition requests;
A batching time management unit that monitors the progress of the set batching time, and a content acquisition request held in the acquisition request holding area corresponding to the batching time when the set batching time is detected. And a cache server control unit that distributes the digital content from the cache server allocated in correspondence with the acquisition request holding area to the client terminal that has made the read content acquisition request. Content distribution control device. Receiving a content acquisition request which is an acquisition request for desired digital content configured by including image or audio data from a client terminal connected via a network;
For each received content acquisition request, assigning a cache server to transmit the requested digital content;
Setting a batching time which is a predetermined waiting time for each digital content to be transmitted by the allocated cache server;
And a step of transmitting the digital content from the cache server to the client terminal in response to elapse of the set batching time. A function of receiving a content acquisition request which is an acquisition request for desired digital content configured by including image or audio data from a client terminal connected via a network;
A function of allocating a cache server for transmitting the requested digital content for each received content acquisition request;
A function of setting a batching time which is a predetermined waiting time for each digital content to be transmitted by the allocated cache server;
A content distribution control program that causes a computer to exhibit a function of transmitting the digital content from the cache server to the client terminal when the set batching time has elapsed. An acquisition request receiving function for receiving a content acquisition request that is an acquisition request for desired digital content configured by including image or audio data from a client terminal connected via a network;
A content detection function for detecting digital content requested in the received content acquisition request;
An acquisition request classification function for classifying the received content acquisition request for each detected digital content;
An acquisition request holding area management function for securing an acquisition request holding area, which is a memory area for temporarily holding the classified content acquisition requests, in a recording medium for each classification;
A cache server setting function for allocating a cache server to which digital content requested by the content acquisition request is to be transmitted for each content acquisition request group held in each of the secured acquisition request holding areas;
A batching time setting function for setting a batching time that is a time for temporarily holding the content acquisition request for each content acquisition request group held in each of the secured acquisition request holding regions;
A batching time management function for monitoring the progress of the set batching time;
When it is detected that the set batching time has elapsed, the content acquisition request held in the acquisition request holding area corresponding to the batching time is read out from the cache server assigned in correspondence with the acquisition request holding area. A cache server control function for delivering the digital content to the client terminal that has made the read content acquisition request;
A content distribution control program for causing a computer to exhibit the above.

Images