JP2019020994A - Network system, node device, cache method, and program - Google Patents

Abstract

To provide a technique capable of avoiding concentration of accesses to a node, suppressing an increase in total communication traffic volume, and efficiently utilizing a cache area, with a simple configuration.SOLUTION: A network system includes a plurality of node devices each transmitting and receiving chunks that are divided files and temporarily holding the chunks. The node device is associated with a group for specifying a chunk to be held by the node device in such a way as to prioritize a group different from other node devices neighboring on the network topology. Chunks are associated with zero or more groups. Further, when receiving a chunk associated with a group assigned to itself, the node device temporarily holds the chunk as a cache in a storage unit.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a network system, a cache method, and a cache program.

  Conventionally, various techniques for intermediately caching files transmitted between computers connected to a network have been proposed. By downloading the cache file instead of the original file, concentration of access to the original file can be avoided and the time until the download is completed on the receiving side can be shortened. For example, when a content file such as a moving image having a relatively large capacity is transmitted / received, a great effect can be expected.

  An autonomous storage device that divides content into a predetermined number of content blocks has also been proposed (for example, Patent Document 1). In this technique, an error correction code block including an error correction code for a content block is generated, and each block of the content block and the error correction code block is distributed and stored in a plurality of autonomous storage devices.

JP 2006-330779 A

  Conventionally, when files are distributed and cached in a plurality of nodes constituting a network system, it generally takes time to determine a file arrangement that can avoid concentration of access and reduce the overall traffic. There was a problem. In addition, for content such as moving images, for example, even in one content, the access frequency may differ for each time interval. In such a case, it is inefficient to cache the entire content.

  The present invention provides a technology for avoiding concentration of access to nodes, suppressing an increase in the total communication volume, and efficiently using a cache area in a network in which content files are transmitted and received, with a simple configuration. The purpose is to provide.

  The network system according to the present invention includes a plurality of node devices that transmit and receive chunks, which are divided files, and temporarily hold the chunks. A group for specifying a chunk to be held by the node device is assigned to the node device so that a group different from other adjacent node devices in the network topology is given priority, and the chunk is a group of zero or more. Associated with. When a node device receives a chunk associated with a group assigned to the node device, the node holding device temporarily stores the chunk as a cache in the storage unit, and the requesting node device requests transmission. If the cache corresponding to the chunk to be held is held, the cache transmission unit that transmits the cache to the requesting node device and the cache corresponding to the chunk requested by the requesting node device are not held A relay unit that requests a chunk from another node.

  In this way, a distributed cooperative system can be realized by distributing node devices belonging to any group on the network. That is, based on the group associated with the chunk, the cache holding unit can easily specify the file that it should hold. In other words, the node device is assigned so that a group different from other adjacent node devices in the network topology is prioritized, thereby avoiding concentration of access to the original file and reducing the overall traffic. In addition, the cache area can be used efficiently.

  Further, the management apparatus further calculates a frequency of requests for the chunks, and the management apparatus holds the calculated high-frequency chunks in more (or more than the same number of) node apparatuses than the low-frequency chunks. You may make it provide the cache management part which changes the group linked | related with a chunk. In this way, even when the trend of the request frequency for the chunk changes, the setting can be changed so that the cache hit rate in the network can be maintained and improved following the change in the trend.

  Further, when the cache transmission unit of the node device transmits the cache held by the storage unit of the node device to the requesting node device, the cache transmission unit counts the number of accesses to the chunk corresponding to the cache, and the number of accesses May be transmitted to the management apparatus. In this way, even when the cache hits in any of the node devices included in the network system, the management device can count the frequency of requests for chunks.

  The file is content data, and the chunk is a file obtained by dividing the content data based on the reproduction time or the data size, and the file may be formed by combining a plurality of chunks. In particular, in the case of content data such as a moving image or audio that is streamed, the more frequently requested data can be cached on the network. For example, if the chunk size is determined based on the data structure or processing unit of content data, the communication efficiency and the data processing efficiency can be improved.

  The network system further includes a terminal that requests content data from the node device in units of chunks, and the terminal has a predetermined frequency of requests among other chunks that constitute a file formed by the requested chunk. You may make it provide the reproduction | regeneration processing part which receives preferentially the chunk more than a threshold value. In this way, it is possible to pre-read and receive a chunk having a high request frequency, and to read it quickly.

  Further, the playback processing unit of the terminal may play back other content before playing back a chunk whose request frequency calculated by the management device is equal to or higher than a predetermined threshold. By inserting other content before a chunk having a high request frequency, the possibility that the user can view other content is increased.

  Further, the playback processing unit of the terminal may combine and play chunks whose request frequency calculated by the management device is equal to or higher than a predetermined threshold among chunks generated by dividing one file. In this way, it is possible to reproduce the content that summarizes one file by using a portion having a high request frequency.

Also, the playback processing unit of the terminal plays back from one of the chunks generated by dividing one file according to the user's operation, the request frequency calculated by the management device being equal to or higher than a predetermined threshold. May be started. In this way, playback can be started from a chunk that is popular, frequently requested, and likely cached in advance.

  The node device according to the present invention transmits and receives chunks that are divided files and temporarily holds the chunks. A group for specifying a chunk to be held by the node device is assigned to the node device, and the file is associated with zero or more groups. In addition, when a node device receives a chunk associated with a group assigned to the node device, a cache holding unit that temporarily holds the chunk as a cache in a storage unit, and the requesting node device requests transmission If the cache corresponding to the chunk to be held is held, the cache transmission unit that transmits the cache to the requesting node device and the cache corresponding to the chunk requested by the requesting node device are not held A relay unit that requests a chunk from another node.

  According to such a node device, a simple configuration in which a group different from other adjacent node devices in the network topology is assigned with priority is avoided, and concentration of access to the original file is avoided, and overall communication is performed. It is possible to reduce the amount and use the cache area efficiently.

  The contents described in the means for solving the problems can be combined as much as possible without departing from the problems and technical ideas of the present invention. The contents of the means for solving the problems can be provided as a device such as a computer or a system including a plurality of devices, a method executed by the computer, or a program executed by the computer. A recording medium that holds the program may be provided.

  In a network in which content files are transmitted and received, it is possible to provide a technique for efficiently using a cache area while avoiding concentration of access to nodes, reducing the total communication amount, and a simple configuration.

It is a figure which shows an example of a structure of a system. It is a figure for demonstrating the information which shows a group attached | subjected to a file. It is a figure which shows an example of the table which matches a chunk and memorize | stores a tag. 3 is a block diagram illustrating an example of a configuration of a terminal 4, a node 3, and a content server 5. FIG. It is a processing flowchart which shows an example of the content delivery process which a network system performs. It is a figure which shows an example of the content of a MPD file. It is a processing flowchart which shows an example of the cache object update process which changes the file of cache object. It is a figure which shows an example of the tag which shows the order | rank based on the access count to the chunk in a predetermined period, and the group allocated in relation to the said chunk. It is a graph which shows the file cached on the mesh type | network network represented by one plane graph, and the number thereof. It is a figure which shows an example of the screen displayed on a terminal. It is a figure which shows an example of the information which shows a popular chunk.

  Hereinafter, embodiments for carrying out the present invention will be described with reference to the drawings. In addition, the structure of embodiment is an illustration and this invention is not limited to the structure shown in embodiment.

<System configuration>
FIG. 1 is a diagram illustrating an example of a configuration of a system according to the embodiment. 1 includes a network 2, nodes 3 (3 a to 3 e) forming the network 2, terminals 4 (4 a to 4 d) connected to the network 2, and a content server 5 connected to the network 2. including. The network 2 is a communication network such as an autonomous system (AS) managed by an ISP (Internet Service Provider). The node 3, the terminal 4, and the content server 5 are computers, and can transmit and receive electronic files to and from other computers connected so as to be communicable.

  In addition, as shown in a large balloon in FIG. 1, the content server 5 stores a content file 6 such as a moving image. Then, in response to a request from the terminal 4, the content server 5 transmits the file 6 to the terminal 4 via the network 2 by, for example, a streaming method. In addition, the file 6 is divided into a plurality of partial files (also referred to as “chunks”) 61 to 63 divided into portions based on time intervals or data sizes. In the present embodiment, as shown in a small balloon in FIG. 1, the file 6 is cached in the node 3 in units of chunks. In the process of relaying the request from the terminal 4 to the content server 5, the node 3 that has received the request transmits the cached chunk to the request source if the node 3 has cached the requested chunk. . If the node 3 that received the request does not cache the requested chunk, the node 3 forwards the request for the chunk to the content server 5 or another node 3.

  In addition, it is assumed that a group for specifying a chunk to be cached in the node 3 is assigned to each node 3 in advance. In other words, each node 3 belongs to one of a plurality of predetermined groups. In the example of FIG. 1, hatching performed in a circle displayed in the vicinity of each node 3 represents a group to which each node 3 belongs. The node 3a belongs to a group of horizontal stripes (referred to as “group A”). The node 3b belongs to a group of vertical stripes (referred to as “group B”). The nodes 3c and 3e belong to a hatched group (referred to as “group C”) from the upper right to the lower left. The node 3d belongs to a hatched group (referred to as “group D”) extending from the upper left to the lower right.

  A certain group may be preferentially assigned a group different from that of the other nodes 3 adjacent in the network topology. That is, groups are assigned to the plurality of nodes 3 so as not to be biased in the network topology. For example, when connecting a new node 3 to the network 2, for example, among a plurality of groups, among other nodes 3 already existing on the network 2, a node 3 with a small number of hops or a node 3 with a short circuit path length. Alternatively, the groups attached to may be excluded in order, and the remaining groups may be assigned to new nodes. In a mesh network or a ring network represented by a plane graph, a so-called four-color theorem may be applied to assign groups so that adjacent nodes do not belong to the same group. In this way, from a certain node 3 to a node 3 to which another group is assigned, the same number of hops can be reached in any group.

Each chunk is associated with zero or more groups that should be cached. In the example of FIG. 1, hatching performed in a rectangle displayed under each of the chunks 61 to 63 represents a group that should hold each chunk. The chunk 61 is assigned four groups A to D. The chunk 62 is assigned with three groups A, B, and D. Two groups A and C are allocated to the chunk 63. There may be a chunk to which no group is assigned. According to such a network system 1, the amount of chunks cached in the network 2 can be adjusted according to the number of groups assigned to the chunks. For example, if more groups are assigned to a frequently accessed chunk, it is possible to increase the probability that the chunk is cached near the requesting node 3.

  In such a system 1, when a node 3 receives a chunk associated with a group assigned to itself, the node 3 caches the chunk. Further, when the requesting node 3 holds a cache corresponding to the chunk that requests transmission, the cache is transmitted to the requesting node 3. On the other hand, if the requesting node device does not hold the cache corresponding to the requested chunk, the request for the chunk is relayed to the other node 3.

FIG. 2 is a diagram for describing information indicating a group attached to a chunk. The file 6 is content data such as a moving image and sound reproduced by the computer. In particular, high-resolution video content, VR (Virtual Reality) content including omnidirectional information, and the like have a large capacity per file and are rarely updated. Therefore, in a CDN (Contents Delivery Network), if a frequently accessed content is targeted for caching, a reduction in traffic due to the cache can be expected.

  For example, information indicating a combination of zero or more of the above-described groups is associated with the file 6 for each chunk. FIG. 2 shows variations of combinations of four types of groups represented by hatching patterns. That is, each group combination has 0 to 4 groups. Further, as shown in the right side of FIG. 2, the combination of groups can be represented by a tag formed by a bit string having the number of digits by the number of groups, for example. The association of the group with the chunk can be stored by holding the tag in association with the identification information of the chunk at each node 3.

  FIG. 3 is a diagram illustrating an example of a table that stores tags in association with chunks. The table in FIG. 3 has attributes of “chunk ID” and “tag”. Chunk identification information is registered in the “chunk ID” field. As the identification information, a name for uniquely identifying the chunk may be used, or information indicating the name of the file and the range of the chunk in the file may be used. The chunk range can be expressed by an offset from the beginning of the file, for example. In the “tag” field, zero or more groups that cache the associated file are registered. Information representing the correspondence between chunks and groups as shown in FIG. 3 (in other words, a list of chunks to be cached) can be stored in each node 3 as a database or other type of file. Also, the information shown in FIG. 3 is centrally managed by a management apparatus such as the content server 5 connected to the network 2, for example, and a list of chunks to be cached is broadcast to the node 3 on the network 2. Each of the nodes 3 may be held, or each of the nodes 3 may be read from a management device on the network 2 at an arbitrary timing.

<Device configuration>
FIG. 4 is a block diagram illustrating an exemplary configuration of the terminal 4, the node 3, and the content server 5. The terminal 4, the node 3, and the content server 5 are general computers.

<Terminal>
The terminal 4 is a computer such as a PC (Personal Computer), a smartphone, or a tablet, and includes a communication I / F 41, a storage device 42, an input / output device 43, a processor 44, and a bus 45.

  The communication I / F 41 is, for example, a network card, and communicates with other computers based on a predetermined protocol. For example, the file 6 is received via the network 2.

  The storage device 42 includes a main storage device such as a RAM (Random Access Memory) and a ROM (Read Only Memory), and an auxiliary storage device (secondary storage) such as an HDD (Hard-disk Drive), an SSD (Solid State Drive), and a flash memory. Device). The main storage device temporarily stores the program and file 6 read by the processor and secures a work area for the processor. The auxiliary storage device stores a program executed by the processor.

  The input / output device 43 is, for example, an input device such as a keyboard and a mouse, an output device such as a monitor, and a user interface such as a touch panel. For example, the user requests the content server 5 to transmit the file 6 via the input / output device 43 or reads the received file 6 to display it.

The processor 44 is an arithmetic processing unit such as a CPU (Central Processing Unit), and performs each process according to the present embodiment by executing an application (also referred to as a program or software). In the example of FIG. 4, functional blocks are shown in the processor 44. For example, the processor 44 functions as the reproduction processing unit 441. For example, based on a user operation, the reproduction processing unit 441 requests transmission of the file 6 via the network 2 or combines the received chunks to reproduce the content.

  The components as described above are connected via the bus 45.

<Node>
The node 3 also includes a communication I / F 31, a storage device 32, an input / output device 33, a processor 34, and a bus 35.

  The communication I / F 31 is, for example, a network card, and communicates with other computers based on a predetermined protocol. For example, the file 6 is transmitted / received between the nodes 3.

  The storage device 32 is a main storage device such as a RAM or a ROM, and an auxiliary storage device such as an HDD, an SSD, or a flash memory. The main storage device temporarily stores programs and data read by the processor and secures a work area for the processor. The auxiliary storage device stores a program executed by the processor, a cache in units of chunks of files to be transmitted and received, information indicating a group assigned to the node 3, and the like.

  The input / output device 33 is, for example, an input device such as a keyboard and a mouse, an output device such as a monitor, and a user interface such as a touch panel.

  The processor 34 is an arithmetic processing unit such as a CPU, and performs each process according to the present embodiment by executing an application. Also in the node 3, functional blocks are shown in the processor 34. For example, the processor 34 functions as each processing unit of the cache holding unit 341, the cache transmission unit 342, and the relay unit 343. The cache holding unit 341 determines whether to hold the received chunk based on the information indicating the group associated with the chunk, and temporarily holds it in the storage device 42. When the cache transmission unit 342 is cached in the storage device 42 of the node 3 in response to a request in units of chunks of a file from another node to the content server, the cached chunk is transferred to the other node 3 or terminal. 4 to send. When the node 3 is located on the route from the terminal 4 to the content server 5, the relay unit 343 transmits a chunk transmission request from another node 3 or the terminal 4 to the content server 5, or a chunk to be transmitted to another node. Relay.

  The components as described above are connected via the bus 35.

<Content server>
The content server 5 such as a content provider also includes a communication I / F 51, a storage device 52, an input / output device 53, a processor 54, and a bus 55.

  The communication I / F 51 is a network card, for example, and communicates with a computer such as the node 3 based on a predetermined protocol. For example, the file 6 is transmitted in response to a request from another node 3 to be linked.

  The storage device 52 is a main storage device such as a RAM or a ROM, and an auxiliary storage device such as an HDD, an SSD, or a flash memory. The main storage device temporarily stores programs and data read by the processor and secures a work area for the processor. It is assumed that the auxiliary storage device stores a program executed by the processor, an original file to be distributed, and the like.

  The input / output device 53 is, for example, an input device such as a keyboard and a mouse, an output device such as a monitor, and a user interface such as a touch panel.

  The processor 54 is an arithmetic processing unit such as a CPU, and performs each process according to the present embodiment by executing an application. Also for the content server, functional blocks are shown in the processor 54. The processor 54 functions as each function unit of the content distribution unit 541 and the cache management unit 542. The content distribution unit 541 reads the file 6 requested to be transmitted from the storage device 52 and transmits the file 6 to the requesting node. The cache management unit 542 determines a chunk to be cached in the node 3 on the network 2 and assigns a group for caching the chunk. For example, the cache management unit 542 may count the number of accesses for each chunk, calculate an access frequency within a predetermined period, and determine a group for caching each file according to the access frequency.

  The components as described above are connected via the bus 55.

<Content distribution processing>
FIG. 5 is a process flow diagram illustrating an example of content distribution processing performed by the network system. The “node” in FIG. 5 corresponds to the node 3 existing on the route from the request source terminal 4 to the content server 5 in FIGS. In other words, there are cases where a plurality of nodes 3 are routed. In FIG. 5, one node is shown for convenience.

  First, the playback processing unit 441 of the terminal 4 requests transmission of content held by the content server 5 (FIG. 5: S1). In the present embodiment, an example will be described in which a content file divided into a plurality of chunks is described by an MPEG-DASH standard MPD (Media Presentation Description) file.

FIG. 6 shows an example of the contents of the MPD file. The MPD file is described in XML (Extensible Markup Language), position information (URL: Uniform Resource Locator) 611, 621, 631,... Indicating the location of the file, and range information 612 indicating the range of each chunk in the file. 622, 632,... That is, the MPD file (content file) 6 includes information on the first chunk 61 including the position information 611 and range information 612, information on the second chunk 62 including the position information 621 and range information 622, and position information 631. And information related to the third chunk 63 including the range information 632. Based on such a file 6, the playback processing unit 441 of the terminal 4 basically requests the content server 5 to transmit a chunk in order starting from the chunk that forms the front range. In one file 6, the above-described chunks may be described for a plurality of files having different image quality and sound quality. In this case, one of a plurality of files is transmitted / received according to the performance of the terminal 4, the state of the communication path, and the like. Further, the present invention can be applied not only to the MPD file but also to a format such as M3U or HLS (HTTP Live Streaming).

  On the other hand, the relay unit 343 of the node 3 receives the request (S2), and determines whether the cache file corresponding to the requested chunk is held in the storage device 32 of the own node (S3). When the cache is held (S3: YES), the cache transmission unit 342 of the node 3 transmits the cache file to the request source (S4). On the other hand, the reproduction processing unit 441 of the requesting terminal 4 receives the cache file and displays the contents on the input / output device 43 such as a monitor (S5).

  When the node 3 does not hold the cache (S3: NO), the relay unit 343 of the relay node relays the request to the adjacent node 3 or the content server 5 (S6). When a request is transferred via a plurality of nodes 3, the same process is repeated. In the present embodiment, the node 3 knows which chunk is cached in the node 3 belonging to which group from the bit position of the tag indicating the group. Therefore, the node 3 stores in advance the group to which the other node 3 connected in the vicinity belongs, so that the node 3 belonging to the group that caches the chunk based on the bit position of the tag of the requested chunk. The request transfer destination may be determined.

  On the other hand, the content distribution unit 541 of the content server 5 receives the transmission request for the chunk (S7), reads the corresponding chunk from the storage device 52, and transmits it (S8). On the other hand, the relay unit 343 of the node 3 relays the received chunk to the request source (S9). Then, the reproduction processing unit 441 of the requesting terminal 4 receives the chunk and displays the content on the input / output device 43 such as a monitor (S10).

  When the chunk is relayed to the request source in S9, the cache holding unit 341 of the node 3 determines whether the chunk is to be cached in the group to which the own node belongs (S11). In this step, it is determined whether information indicating a group to which the own node belongs is associated with the chunk. For example, the storage device 32 of the node 3 holds information as shown in FIG.

  When it is determined that the relayed chunk is a cache target (S11: YES), the cache holding unit 341 of the node 3 temporarily holds the chunk in the storage device 32. On the other hand, when it is determined that the relayed chunk is not a target to be cached, the node 3 does not hold the chunk. The network system 1 repeats the file distribution process as shown in FIG. 5 in response to the file transmission request from the request source node.

<Effect>
In the present embodiment, the nodes are grouped so that the same group is not biased in the network 2. In other words, nodes that cache the same chunk are distributed on the network by a simple method in which a group different from other adjacent node devices in the network topology is preferentially allocated, and the nodes are located close to any node to some extent. A sub-optimal cache arrangement in which there are cache servers belonging to each group. Further, as shown in FIG. 1, information indicating a plurality of groups may be associated with one chunk. If the more frequently accessed chunks are cached by nodes belonging to many groups, the cache hit rate on the network can be improved. In particular, in the case of a content file such as a moving image, the access frequency of the beginning portion is high, and the user may stop viewing or skip a portion thereafter. In addition, the access frequency of a part of the moving image may be increased, for example, when a link to the moving image specifying the position to start reproduction is posted to SNS (Social Networking Service). According to the present embodiment, the more frequently accessed chunks can be cached, the increase in traffic in the entire network system can be suppressed. Even when the requested frequency differs in units of chunks, the cache area can be efficiently utilized by caching the chunks having a high request frequency.

<Cache target decision processing>
In the network system described above, the priority of the cache for the chunk may be determined, and the information indicating the group associated with the chunk (in other words, the list of chunks to be cached) may be dynamically changed. The priority of the cache is determined based on, for example, the access frequency to each chunk calculated by the cache management unit 542 of the content server 5. The content server 5 is an example of an “access management device” according to the present invention that calculates the frequency of access to a chunk and determines a chunk to be cached. Note that an access management device may be provided in any one of the networks 2 separately from the content server 5.

  FIG. 7 is a process flow diagram illustrating an example of a cache target update process for changing a cache target chunk. For example, the network system may perform processing as shown in FIG. 7 in addition to the content distribution processing shown in FIG. In FIG. 7, the processes corresponding to those in FIG.

  First, the node 3 counts the number of times the cache is transmitted in response to the request, and transmits the counted number to the content server 5 at a predetermined timing (FIG. 7: S21). On the other hand, the content server 5 receives the number of cache transmissions from the node 3 and calculates the number of accesses for each chunk (S22). In this step, the access log stored in the node 3 may be received and the number of accesses for each chunk may be calculated.

  Further, for example, in the step of transmitting a chunk in S8 of FIG. 5, the content server 5 further writes an access log to the chunk in the storage device 32 (FIG. 7: S23). The access frequency for each chunk can be calculated by counting the access log stored in this step and the number of accesses received in S22 for each chunk. In this step, the cache management unit 542 of the content server 5 totals the number of accesses for each chunk (that is, access frequency) in a predetermined period at a predetermined timing.

  Thereafter, the cache management unit 542 of the content server 5 causes a larger number of groups to be cached at a predetermined timing, assuming that the more frequently used chunks have higher cache priority. That is, information indicating more groups is associated with the chunk. Further, for example, the cache target is updated when a condition such as a predetermined operation time has arrived or a predetermined period has elapsed since the previous process is satisfied. Then, the cache management unit 542 calculates an access frequency for each chunk and determines a cache target (S24).

FIG. 8 is a diagram illustrating an example of a tag indicating a rank based on the number of accesses to a chunk during a predetermined period and a group assigned in association with the chunk. Content server 5
8 summarizes a log of chunks transmitted by the content server 5 in response to a request and a log indicating that the node 3 has transmitted a cache for a predetermined period, and generates statistical data for each chunk as shown in FIG. . The table shown in FIG. 9 includes items of “rank”, “number of groups”, and “tag”. In the “rank” field, a rank range determined based on the access frequency to the chunk in a predetermined period is recorded. In S24 of FIG. 7, the order in a predetermined period is determined. The “number of groups” field indicates the number of groups that cache the chunks of each rank. In the “tag” field, information indicating a group for caching each rank chunk is recorded in the tag format described above.

  Further, the cache management unit 542 determines a group for caching each chunk according to the access frequency (FIG. 7: S25). In this step, a cache group is assigned to each chunk. First, the degree of popularity shown in FIG. 8 is determined based on the number of accesses during a predetermined period. For example, when the node 3 is divided into four groups, the degree of popularity is classified into five levels. Also, the most popular chunk is cached in all four groups, the second most popular chunk is cached in three groups, and the third most popular chunk is Cache in the group and chunks belonging to the fourth most popular are cached in one group. Note that chunks belonging to the least popular level are not cached in any group.

  In this way, the number of groups of nodes that cache the chunk is determined according to the degree of popularity, and thereafter, a combination of groups is cyclically assigned to chunks to which the same number of groups are assigned. For example, tags determined by cyclically assigning combinations of three groups are recorded in the “tag” field of the records with ranks 21 to 100 in FIG. In S25, the tag associated with the chunk is updated based on the access frequency to each chunk in the most recent predetermined period.

  Then, the cache management unit 542 distributes a list indicating the groups for which each chunk is cached to each node (FIG. 8: S26). In this step, for example, the information shown in the “chunk name” and “tag” columns in FIG. 8 is broadcast to each node 3. When the node 3 receives the list, the node 3 holds the list in the storage device 32, and thereafter determines a target chunk to be cached by the own node based on the list (S27). Based on the information indicating the group of such a list, each node can determine the destination to request the chunk. Further, when the list is updated, an unnecessary cache in the node 3 may be deleted.

<Effect>
According to the cache target update process, the file distributed from the content server 5 can be cached in units of chunks according to the number of accesses. Even when the access frequency trend changes, the setting can be changed so that the cache hit rate in the network can be maintained and improved following the change in the trend. Note that the management apparatus that calculates the access frequency and updates the list of chunks to be cached may be performed by the content server 5 or may be performed by another apparatus connected to the network 2. Further, the access log for each chunk stored in each node 3 may be held in a predetermined device on the network so that the content server 5 or the management device can read the access log in each relay node.

<Files to be cached and the number of nodes that cache the files>
In the present embodiment, chunks with high access frequency are targeted for caching, and the higher the access frequency, the more nodes are cached in the network. The number of nodes that cache chunks can be arbitrarily set by the user. Note that the number of nodes that cache chunks is determined by the number of nodes that belong to each group connected to the network and the number of groups that hold the chunks that are set in association with the chunks. The number of nodes that cache the chunks may be determined according to a sub-optimal cache arrangement obtained using, for example, a genetic algorithm.

FIG. 9 is a graph showing files (chunks) cached on the mesh network represented by one plane graph and the number thereof. The horizontal axis in FIG. 9 represents the serial numbers of files sorted in descending order of access frequency, and the vertical axis represents the number of nodes that cache each file. The setting of a tag attached to a file may be determined according to the type of file cached on a node and the number of nodes that cache the same file in a cache arrangement obtained using a genetic algorithm. The broken line in FIG. 9 represents the relationship between a file in a sub-optimal cache arrangement obtained by a genetic algorithm (GA) and the number of files to be cached on the network in a mesh type network. . A solid line is an example in which the network system according to the present embodiment approximates the number of cache files in the network obtained by the genetic algorithm. In the present embodiment, the number of nodes on the network that caches the file can be adjusted by changing the group assigned to each file, which is represented by the tag described above. Further, by making the number of types of groups larger than 4, it is possible to improve the accuracy of approximation for the sub-optimal arrangement. Even in a hierarchically connected network according to the present embodiment, a plurality of files that are accessed at a predetermined frequency are defined, and the above files are set to approximate values of a sub-optimal cache arrangement obtained by some algorithm. It is also possible to determine at least one of a group to be assigned to and a group to be assigned to a node on the network.

  For example, as shown in FIG. 9, the number of groups that cache chunks is determined for each degree of popularity, such as within 20th place, 21-100th place, 101st-250th place, etc., based on the access frequency. Then, tags associated with the degree of popularity are cyclically assigned in order from the top. The tag is described by a 4-bit bit vector indicating a group that should hold the chunk. The bit vector is formed by connecting digits representing groups in each layer in a predetermined order. In FIG. 8, each of the 4 bits corresponds to 4 groups, and indicates whether or not a chunk is cached in each group.

<Modification 1>
FIG. 10 is a diagram illustrating an example of a screen displayed on the input / output device 43 of the terminal 4. In this modification, it is possible to find a chunk whose access frequency is equal to or higher than a predetermined threshold (that is, popular). For example, by adding information indicating that the access frequency for each chunk or a popular chunk is added to the MPD file shown in FIG. 6, the information indicating that the access frequency is equal to or higher than a predetermined threshold or is a popular chunk is obtained. A playback position may be skipped to a predetermined chunk that is added, or a file containing information indicating the access frequency and popularity of each chunk is prepared separately from the MPD file, and the MPD file is combined with the predetermined chunk. The playback position may be skipped to the chunk. Note that the playback position is not necessarily the beginning of the chunk, and an arbitrary position in the content may be specified.

FIG. 11 is a diagram illustrating an example of information indicating popular chunks generated separately from the MPD file. The table of FIG. 11 includes attributes of “chunk ID” and “reproduction position”. In the “chunk ID” field, identification information for specifying a chunk is registered. With this identification information, the tag for each chunk shown in FIG. 3, the chunk of the MPD file shown in FIG. 6, and the reproduction position in FIG. 11 are linked. In the “playback position” field, the position where playback determined based on the access frequency is to be started is registered as information representing the time from the beginning of the content. Information as illustrated in FIG. 11 may be distributed to the terminal 4, and the reproduction processing unit 441 of the terminal 4 may generate a screen as illustrated in FIG. 10 using the distributed information.

  The screen shown in FIG. 10 is generated by an application that reproduces content executed on the terminal 4, for example. The screen of FIG. 10 includes a moving image display area 411, a seek bar 412, a forward skip button 413, a backward skip button 414, and a scene list 415. In the area 411, a moving image is displayed. The seek bar 412 in FIG. 10 indicates the current playback position in the whole as a triangle, and indicates a part with high access frequency as a star. This star mark corresponds to the reproduction position shown in FIG. Then, the forward skip button 413 and the backward skip button 414 cause the playback position to be skipped to a frequently accessed part based on the time described in FIG. In the scene list 415, a list of frequently accessed parts is displayed. The scenes displayed in the list 415 also correspond to the playback positions shown in FIG. When a scene in the list 415 is selected by a user operation, a moving image can be reproduced from each scene based on the time described in FIG.

  As described above, by providing the skip buttons 413 and 414 or the list 415, it is possible to start reproduction from a chunk whose access frequency is equal to or higher than a threshold according to a user operation. That is, the playback position can be moved to a chunk that many users are viewing. Therefore, it is possible to reproduce from scenes with high access frequency such as a score scene of sports. Further, by caching chunks corresponding to scenes with high access frequency in more nodes 3, it is possible to suppress an increase in traffic on the network. In other words, traffic can be reduced by suggesting the user to replay from a previously cached chunk.

<Modification 2>
Alternatively, a digest of content may be generated by extracting and combining chunks having an access frequency equal to or higher than a predetermined threshold. That is, among the chunks generated by dividing one file, the chunks whose access frequency calculated by the management device is equal to or higher than a predetermined threshold are combined and played back. Such processing may be realized, for example, by the content server 5 or the like generating an MPD file including only chunks whose access frequency is equal to or higher than a predetermined threshold as a content digest, or content provided in the terminal 4 The reproduction program may be realized by requesting only chunks whose access frequency is equal to or higher than a predetermined threshold among chunks described in the MPD file. That is, the process of forming the original file by combining a plurality of chunks may be performed by the node 3 or the terminal 4.

  By caching chunks corresponding to scenes with high access frequency in more nodes 3, an increase in traffic on the network can also be suppressed according to this modification.

<Modification 3>
Further, the playback processing unit 441 of the terminal 4 may receive (prefetch) preferentially a chunk whose access frequency is equal to or higher than a predetermined threshold. In this way, the terminal 4 can pre-read and download the part that the user wants to view, and if the user performs an operation to skip to the scene, playback can be started immediately. Further, the cache transmission unit 342 of the node 3 may be the main body, and the chunks whose access frequency is equal to or higher than a predetermined threshold may be push delivered to the terminal 4. In any case, the playback processing unit 441 of the terminal 4 can preferentially receive a chunk whose access frequency is equal to or higher than a predetermined threshold among the chunks constituting the file. Then, when the chunk is played back, the chunk can be held in the storage device 42 of the terminal 4.

<Modification 4>
Moreover, you may make it reproduce | regenerate by inserting other content like an advertisement before the chunk whose access frequency is more than a predetermined threshold among one content. In such a process, for example, the reproduction processing unit 441 of the terminal 4 cooperates with the content server 5 or the node 3 to reproduce another content by interrupting a chunk whose access frequency is a predetermined threshold or more. Also, other contents may be cached in advance in the node 3. By inserting other content before the popular chunk in this way, the possibility that the user can view other content is increased. The insertion of other contents according to this modification may be push-distributed to the terminal 4 mainly by the cache transmission unit 342 of the node 3.

<Modification 5>
Further, the content data is not limited to moving images and audio, and may be VR (Virtual Reality) moving images including a three-dimensional object, for example. At this time, a file format in which transmission / reception can be controlled in units of objects in the VR moving image may be adopted so that not only the time-divided chunks are cached but also the three-dimensional objects can be cached. That is, a three-dimensional object that is directed to many users and is often a display target is cached in a node 3 belonging to a larger number of groups, is not noticed by the user, and is less likely to be displayed. The object hardly caches on the network 2. In addition, 3D objects with different quality in stages are prepared in advance, high quality 3D objects that are often displayed are cached, and 3D objects that are rarely displayed are You may make it cache a low quality thing. Also according to such a modification, an increase in traffic on the network can be suppressed, and a storage area for holding a cache can be effectively used.

  That is, when handling content data with a special data structure or processing unit, such as a data structure file in which three-dimensional objects, which are data processing units, are stored together, chunks are created based on the data structure or processing unit. The data size may be determined. In this way, it is possible to improve communication efficiency and data processing efficiency when files are transmitted / received and cached in units of chunks.

In addition, the cache method according to the embodiment or the modification may be applied to a moving image in a file format in which playback quality and content data amount can be adjusted by combining layers having different image quality as content data. For example, H.M. H.264 SVC (Scalable Video Coding) prepares a video file that includes layers with different image quality in stages, and the higher the access frequency, the higher the image quality according to the access frequency to the video for each time interval. You may make it cache many layers. Also, a large number of low image quality layers may be cached regardless of the access frequency. In this way, at least a low image quality layer can be delivered to the terminal 4 at an early stage regardless of the congestion status of the line.

  The matters described in the above embodiments and modifications can be implemented in appropriate combination.

<Others>
The present invention includes a computer program that executes the above-described processing. Furthermore, a computer-readable recording medium on which the program is recorded also belongs to the category of the present invention. With respect to the recording medium on which the program is recorded, the above-described processing can be performed by causing the computer to read and execute the program on the recording medium.

  Here, the computer-readable recording medium refers to a recording medium in which information such as data and programs is accumulated by electrical, magnetic, optical, mechanical, or chemical action and can be read from a computer. Examples of such a recording medium that can be removed from the computer include a flexible disk, a magneto-optical disk, an optical disk, a magnetic tape, and a memory card. Further, examples of the recording medium fixed to the computer include a hard disk drive and a ROM.

1: Network system 2: Network 3: Node 31: Communication I / F
32: Storage device 33: Input / output device 34: Processor 341: Cache holding unit 342: Cache transmission unit 343: Relay unit 35: Bus 4: Terminal 41: Communication I / F
42: storage device 43: input / output device 44: processor 441: reproduction processing unit 45: bus 5: content server 51: communication I / F
52: storage device 53: input / output device 54: processor 541: content distribution unit 542: cache management unit 55: bus

Claims (11)

A network system including a plurality of node devices that transmit and receive chunks that are divided files and temporarily hold the chunks,
A group for specifying a chunk to be held by the node device is assigned to the node device so that a group different from other adjacent node devices in the network topology is given priority.
The chunk is associated with zero or more of the groups;
The node device is
When the chunk associated with the group assigned to itself is received, a cache holding unit that temporarily holds the chunk as a cache in a storage unit;
When the requesting node device holds the cache corresponding to the chunk that requests transmission, a cache transmission unit that transmits the cache to the requesting node device;
When the cache corresponding to the chunk requested by the requesting node device is not held, a relay unit that requests the chunk from another node;
A network system comprising: A management device for calculating the frequency of requests for the chunk;
The said management apparatus is provided with the cache management part which changes the said group linked | related with the said chunk so that the calculated chunk with high frequency may be hold | maintained in more node apparatuses than the chunk with low frequency. Network system. When the cache transmission unit of the node device transmits the cache held by the storage unit of the node device to the requesting node device, the cache transmission unit counts the number of accesses to the chunk corresponding to the cache, The network system according to claim 2, wherein the number of accesses is transmitted to the management device. The file is content data;
The network system according to claim 2 or 3, wherein the chunk is a file in which the content data is divided based on a reproduction time or a data size, and the file is formed by combining a plurality of chunks. The network system further includes a terminal that requests the content data in chunk units from the node device,
The network according to claim 4, wherein the terminal includes a reproduction processing unit that preferentially receives a chunk whose request frequency is equal to or higher than a predetermined threshold among other chunks that form a file formed by the requested chunk. system. The network system according to claim 5, wherein the reproduction processing unit of the terminal reproduces other content before reproducing a chunk whose request frequency calculated by the management device is equal to or greater than a predetermined threshold. The reproduction processing unit of the terminal combines and reproduces chunks that are generated by dividing one file and whose request frequency calculated by the management device is equal to or higher than a predetermined threshold. 6. The network system according to 6. The playback processing unit of the terminal is any one of chunks generated by dividing one file according to a user's operation, and the frequency of the request calculated by the management device is equal to or higher than a predetermined threshold. The network system according to any one of Claims 5 to 7, wherein reproduction is started from. A node device that transmits and receives chunks that are divided files and temporarily holds the chunks,
The node device is assigned a group for specifying a chunk to be held by the node device,
The file is associated with zero or more of the groups;
The node device is
When the chunk associated with the group assigned to itself is received, a cache holding unit that temporarily holds the chunk as a cache in a storage unit;
When the requesting node device holds the cache corresponding to the chunk that requests transmission, a cache transmission unit that transmits the cache to the requesting node device;
When the cache corresponding to the chunk requested by the requesting node device is not held, a relay unit that requests the chunk from another node;
A node device comprising: A cache method executed by a node device that transmits and receives chunks that are divided files and temporarily holds the chunks,
The node device is assigned a group for specifying a chunk to be held by the node device,
The file is associated with zero or more of the groups;
The node device is
When the chunk associated with the group assigned to itself is received, the chunk is temporarily held in the storage unit as a cache,
When the requesting node device holds the cache corresponding to the chunk that requests transmission, the cache is transmitted to the requesting node device,
A cache method for requesting the chunk from another node when the cache corresponding to the chunk requested by the requesting node device is not held. A program that is executed by a node device that transmits and receives chunks that are divided files and temporarily holds the chunks,
The node device is assigned a group for specifying a chunk to be held by the node device,
The file is associated with zero or more of the groups;
In the node device,
When the chunk associated with the group assigned to itself is received, the chunk is temporarily held in the storage unit as a cache,
When the requesting node device holds the cache corresponding to the chunk that requests transmission, the cache device is transmitted to the requesting node device,
A program for causing another node to request the chunk when the cache corresponding to the chunk requested by the requesting node device is not held.

Images