KR101128292B1 - System and method for content delivery using multiple cache server, and cache server thereof - Google Patents

Abstract

The present invention relates to a content delivery system and method using a plurality of cache servers, and a cache server, the storage area of the data fragment storage included in the plurality of cache servers by receiving the content data, divided into pieces of a certain size Is stored in each partitioned area with the corresponding schedule size, and upon receiving a request from the client, the number of pieces to be included in the payload of one data packet is determined and the stored pieces are stored in the order of the pieces corresponding to the content data. Therefore, it is sequentially loaded into the memory and transmitted to the client, thereby minimizing the fragmentation loss due to incomplete transmission of contents, enabling the appropriate amount of data transmission according to network traffic conditions, and increasing the data transmission speed. .

CDN, Cache Server, Replication Server, Ghost Server

Description

System and method for content delivery using multiple cache servers, and cache server TECHNICAL FIELD

The present invention relates to a content delivery system and method, and a cache server thereof. Particularly, a cache server divides content data into pieces of a predetermined size and stores them in each partition of a plurality of cache servers divided by a corresponding size, and requests a client's request. The present invention relates to a content delivery system and method using a plurality of cache servers for retrieving and storing divided fragments into memory, and a cache server thereof.

In general, the number of contents (eg, videos, music, files, etc.) that a content provider (CP) provides to a client is very large. There is a heavy load for a web server to provide some or all of these to each of the multiple clients one-to-one. This is because a web server is connected to a single Internet Service Provider (ISP), whereas clients requesting content exist in multiple ISPs.

Therefore, traffic between ISPs increases exponentially as the number of content providers and clients increases, and transmission delay and / or transmission failure rate due to traffic congestion increases. In the worst case, the web server may fail over.

The CDN (Content Delivery Network) has been introduced to overcome this problem. A CDN is a collection of content that is concentrated on a content provider's web server that is remote from clients, such as pictures, banners, video, or audio. Point-of-presence is stored in the host server which is widely distributed, and the client connects to the host server through the returned path by returning the path of the host server closest to the user when requested by the client. This service allows you to receive content. The host server is variously referred to as a cache server, a replication server, or a ghost server (hereinafter referred to as a cache server).

Conventionally, when a cache server transmits a specific content to a client, the cache server refers to a storage address of the content stored therein and loads it into a memory, and divides it into a predetermined size and transmits the content. In this case, the entire file must be loaded into the memory regardless of the size of the content data, and when the client stops the content request at the time when the transfer of the entire file is not completed, a splitting loss is caused. In addition, there is a problem that efficient network management is difficult by transmitting data packets of a certain size regardless of network traffic.

The present invention has been proposed to solve the above-described problems, and the cache server divides the content data into pieces of a certain size, stores them in each partition of the plurality of cache servers divided by the corresponding size, and stores the data in real time. Content delivery system and method using a plurality of cache server that can determine the number of fragments included in the packet to be loaded into the cache server memory receiving the content request from the client and to sequentially transmit the number of pieces to the client by the corresponding number, The purpose is to provide a cache server.

Content delivery method using a plurality of cache server of the present invention for achieving the above object, the cache server receives the content data from the content server, divided into pieces of a certain size, the cache server data divided into the predetermined size A storage step of distributing and storing each partition of the fragment storage and one or more other cache server data in the fragment storage; fragments to be included in one data packet when the cache server receives a request for the content data from a client A determination step of determining the number of pieces, and each piece of the pieces stored in each partition of the cache server data piece storage and pieces stored in each partition of the other cache server data piece storage each piece corresponding to the content data. Sequentially lit into internal memory At the same time as the rudder, the step of transmitting to the client including the retrieved fragment in the payload of the data packet by the number of the determined pieces; It includes.

In the content transmission method using a plurality of cache servers of the present invention, the storing step is characterized in that for storing the start address information of each divided area for storing the divided pieces in the corresponding memory.

In the content transmission method using a plurality of cache servers of the present invention, the transmitting step is characterized in that the stored fragments are loaded into the memory with reference to the start address information of each division area.

In the content transmission method using a plurality of cache servers of the present invention, the predetermined size is characterized in that the maximum transmission unit divided by a specific integer.

In the content transmission method using a plurality of cache servers of the present invention, the determining step is characterized in that the number of the pieces are determined in consideration of network traffic.

In the content transmission method using a plurality of cache servers of the present invention, the determining step, characterized in that the number of the pieces are determined so that the value multiplied by the number of pieces and the predetermined size does not exceed the maximum transmission unit. .

In the content transmission method using a plurality of cache servers of the present invention, the transmitting step, the cache server makes a request for the pieces stored in each of the other cache server to the other cache server, the receiving the request Each of the other cache servers loads the fragments stored in each partition of the corresponding data fragment storage into the memory included therein and simultaneously transmits the fragments to the cache server, and the cache server receives the fragments transmitted from the other cache servers. To be loaded into the memory of the cache server.

In the content transmission method using a plurality of cache servers of the present invention, the transmitting step is characterized in that the number of transmissions for the transmitted fragments at the same time as the transmission.

In the content transmission method using a plurality of cache servers of the present invention, the transmitting step, characterized in that the remaining pieces except for the pieces whose transmission frequency exceeds the threshold value as a result of the counting is characterized in that deleted from the memory.

In the content transmission method using a plurality of cache servers of the present invention, the cache server is included in one data packet when receiving the request for the content data from any client after deleting the remaining pieces from the memory. The number of fragments is re-determined, and the fragments stored in each division region of the cache server data fragment storage and the fragments stored in each division region of the other cache server data fragment storage among the remaining remaining fragments correspond to the content data. At the same time, the fragments existing in the memory of the cache server and the remaining fragments are reloaded in the payload of the data packet by the number of fragments determined again, in the order of the fragments. To send to any client, including .

In the content transmission method using a plurality of cache servers of the present invention, the transmission step, the remaining pieces except the pieces stored in each partition area of the cache server data piece storage of the pieces of the counted number of transmissions exceeded a threshold value; Deletes from the memory inside the cache server and sends a message to the other cache server so as not to delete, from the memory inside the other cache server, a piece imported from the other cache server among the pieces whose transmission frequency exceeds a threshold. Characterized by request.

In the content transmission method using a plurality of cache server of the present invention, the cache server, if the request for the content data from any client after deleting the remaining pieces from the memory and making the request, one data The number of fragments to be included in the packet is determined again, and among the remaining fragments, fragments stored in each partition of the cache server data fragment storage, fragments present in memory of the other cache server, and internally, the other cache server. Among the fragments that do not exist in the memory of the fragments stored in each partition of the other cache server data fragment storage is sequentially loaded back into the memory of the cache server in the order of each fragment corresponding to the content data, The payload of the data packet is determined by the number of pieces determined again. And a fragment present in the cache server's memory and the remaining fragments re-loaded to the arbitrary client.

Content delivery system using a plurality of cache server of the present invention for achieving the above object, the content server for providing the content, and when receiving the content data from the content server is divided into pieces of a certain size to the It is distributed and stored in each division of the internal data fragment storage divided into a certain size and each division of the data fragment storage of one or more other cache servers, and is included in one data packet when a request for the content data is received from a client. Determine the number of pieces to be sliced, and sequentially store pieces stored in each partition of the cache server data piece storage and pieces stored in each partition of the other cache server data piece storage according to the order of each piece corresponding to the content data. Into the internal memory at the same time, By the number of pieces specified by including the pieces which are loaded on the payload of the data packet includes a cache server that is transmitted to the client.

The cache server of the present invention for achieving the above object, through the input and output unit for transmitting and receiving data, a data fragment storage for storing a piece of content data of a predetermined size in the divided area divided into the predetermined size, and through the input and output unit Upon receiving the content data, the content data is divided into pieces of the predetermined size and distributed and stored in each partition of the internal data fragment storage and at least one other cache server data fragment storage. Upon receiving the request, the number of fragments to be included in one data packet is determined, and the fragments stored in each division of the data fragment storage and the fragments stored in each division of the other cache server data fragment storage are stored in the content data. Sequentially in the order of the corresponding pieces And a control module which loads the loaded fragments into the payload of the data packet by the number of the determined fragments and transmits them to the client through the input / output unit.

In the cache server of the present invention, the control module is configured to store start address information of each division area for storing the divided pieces in a corresponding memory.

In the cache server of the present invention, the transmitting step refers to loading the stored fragment into the memory by referring to the start address information of each division area.

In the cache server of the present invention, the control module makes a request for a fragment stored by each of the other cache servers to the other cache server through the input / output unit, and each of the other cache servers receiving the request is a data fragment. The pieces stored in each division area of the storage are loaded into the memory included therein and transmitted to the input / output unit, and the control module receives the pieces transmitted from the other cache server through the input / output unit to receive the pieces of memory in the control module. It is characterized by a call to.

In the cache server of the present invention, the control module is characterized in that the number of transmissions for the transmitted fragments at the same time as the transmission.

In the cache server of the present invention, the control module may delete the remaining pieces from the memory except for the pieces whose transmission count exceeds the threshold.

In the cache server of the present invention, the control module, after deleting the remaining pieces from the memory, the number of pieces to be included in one data packet when receiving a request for the content data from any client through the input and output unit; Re-determining, and in accordance with the order of each piece corresponding to the content data, the pieces stored in each divided area of the data piece storage and the pieces stored in each divided area of the other cache server data piece storage among the remaining pieces are deleted. At the same time, re-loading into the memory of the control module at the same time, the input and output including the fragment present in the memory of the control module and the remaining re-loaded pieces in the payload of the data packet by the number of the pieces determined again Sent to the arbitrary client via .

In the cache server of the present invention, the control module deletes, from the memory inside the control module, the remaining pieces, except for the pieces stored in each division area of the data piece storage, among the pieces whose transmission count exceeds the threshold. And transmitting a message to the other cache server so as not to delete the fragment retrieved from the other cache server among the pieces whose transmission frequency exceeds a threshold from the memory inside the other cache server.

In the cache server of the present invention, the control module deletes the remaining pieces from the memory and makes a request and receives the request for the content data from any client through the input / output unit in one data packet. The number of fragments to be re-determined and pieces of the remaining fragments stored in each partition of the data fragment storage, fragments existing in the memory inside the other cache server, and fragments existing in the memory inside the other cache server are not present. Among the fragments, the fragments stored in each division area of the other cache server data fragment storage are sequentially loaded back into the memory inside the control module according to the order of the respective fragments corresponding to the content data, and at the same time, In the control module in the payload of the data packet by the number Including fragments and pieces, which are the rest of the back up to the present in the memory through the input and output characterized by transmitting to the arbitrary client.

According to the present invention, the cache server loads a certain number of pieces of content stored in the data store into the memory and transmits the pieces of content to the client, thereby minimizing the splitting loss that occurs when the content transmission is incompletely terminated. You can select and transfer the amount of data, and transfer the content data at a faster rate than bringing the content data into memory from the data fragment store of a single cache server and sending the divided data pieces into divided sizes. It can be stored in a repository, quickly loaded into memory, and sent, and a piece of memory that is requested more than other pieces can be stored in memory and quickly delivered to the client.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. However, in describing in detail the operating principle of the preferred embodiment of the present invention, if it is determined that the detailed description of the related known functions or configurations may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted. In addition, the same reference numerals are used for parts having similar functions and functions throughout the drawings.

1 is a block diagram of a content delivery system according to an embodiment of the present invention. Referring to FIG. 1, the content providing system according to the present embodiment includes a first cache server to n-th cache servers 100-1 to 100-n, a content server 200, and a client capable of transmitting and receiving data through a network 400. And 300.

The network 400 refers to a communication network for transmitting and receiving data between the first cache server to the nth cache server 100-1 to 100-n, the content server 200, and the client 300.

The client 300 refers to a computer that receives information from a server connected to the network 400. When the client 300 inputs a Uniform Resource Locater (URL) for requesting specific content into a browser, the client 300 accesses a name server such as a domain name server (DNS) to provide a corresponding content for a domain name. It converts to the IP address of one of the content server 200 or the first cache server to the n-th cache server (100-1 to 100-n) and the server corresponding to the IP address.

The content server 200 is a server that stores web pages and objects provided by the content provider and responds to content requests from the client 300. That is, the content server 200 serves to transmit the content requested from the client 300, and some of the contents held by the content server 200 are first cache servers to nth cache servers 100-1 to 100-. n) to be transmitted to and stored in one of the cache server. Through this, it is possible to quickly provide the content according to the request of the client 300 in one cache server of the first cache server to n-th cache server (100-1 to 100-n).

The first cache server to the n-th cache server (100-1 to 100-n), for example, is called a replication server (ghost server) or a ghost server (ghost server), the content server 200 is It stores a copy of the content data to be stored and managed to provide a prompt when there is a content request from the client (300).

In the present invention, when the first cache server 100-1 receives the content data from the content server 200, the first cache server 100-1 is divided into pieces of a predetermined size and the data pieces of the first cache server 100-1 divided into the corresponding sizes. The storage area is divided and stored in each of the division areas of the data fragment storage of the second cache server to the n-th cache server (100-2 to 100-n). When the first cache server 100-1 stores the fragmented data fragments, the first cache server 100-1 or the first cache server to the n-th cache server store the start address information of each partition in which the fragments are stored. (100-1 to 100-n) may be stored in an internal memory, so that each partition area of the data fragment storage may be loaded into the memory with reference to corresponding start address information.

At this time, the first cache server 100-1 divides the content data into pieces of a predetermined size, and, for example, the first cache server to the nth cache server 100-1 to 100- n) divided by 'n', the first fragment having the remainder '1' is stored in the data fragment storage of the first cache server 100-1, and the remaining fragments of the '2' through '0'. The data fragment storage of the second cache server 100-2 to the nth cache server 100-n may be transmitted to each of the distributed storage.

In this case, the predetermined size for dividing the content data may be, for example, a size in which a maximum transfer unit (MTU) is obtained as a specific integer. The maximum transmission unit refers to the maximum amount of data that can be transmitted in a packet or frame based network, such as a Transmission Control Protocol / Internet Protocol (TCP / IP) network.

Thereafter, when the first cache server 100-1 receives a request for content data from the client 300, the first cache server 100-1 determines the number of pieces to be transmitted, including one data packet transmitted to the client 300. In this case, the number of pieces determined may be, for example, a relatively small value when the network 400 traffic is high, and conversely, a relatively large value when the network 400 traffic is low.

In determining the number of pieces to be transmitted by the first cache server 100-1 in one data packet, for example, the product of the determined number of pieces and a predetermined size used to divide the data is the maximum transmission unit. It can be determined not to exceed. This is because when the amount of data included in one data packet exceeds the maximum transmission unit, the data packet must be divided and routed again by a router existing on the network 400, which may cause a delay in transmission time.

After the first cache server 100-1 determines the number of pieces, the pieces of content data stored in each division area of the data piece storage are sequentially loaded into the memory according to the order of each piece corresponding to the content data. That is, the first cache server 100-1 is a fragment stored in each division area of the first cache server 100-1 data fragment storage and the second cache server to n-th cache server 100-2 to 100-n. The fragments stored in each division area of the data fragment storage are sequentially loaded into the memory of the first cache server 100-1 in consideration of the order in which the content data is divided. In this case, as described above, it can be loaded into the memory by referring to the start address information of each partition in which each fragment is stored, thereby eliminating the need for file I / O (Input / Output) operation like a general cache server. It is possible to transmit to the client 300 at high speed.

At this time, the process of loading each piece into the memory, for example, in the case of the pieces stored in each partition of the second cache server to the n-th cache server (100-2 to 100-n) data fragment storage, the first cache The server 100-1 stores the second cache server through the nth cache server 100-2 through 100-n in pieces stored by each of the second cache server through the nth cache server 100-2 through 100-n. The second cache server to the n-th cache server (100-2 to 100-n) receiving the request for the transmission of the pieces stored in each partition of the data fragment storage in the memory contained therein at the same time The first cache server 100-1 may be transmitted in a manner of transmitting.

In this manner, all fragments are stored in the data fragment storage of the first cache server 100-1, and the response speed is faster than that of the fragments stored in the memory. In other words, it takes a long time to store all data fragments in one data fragment storage and load them into memory sequentially by referring to each fragment's storage address. Recalling a piece in the memory of each cache server, receiving it from one cache server and transmitting to the client 300 is faster response.

Each time the cache server 100-1 receiving each fragment receives the determined number of pieces of data, the cache server 100-1 sequentially includes the fragments in the payload of the data packet and transmits the fragments to the client 300.

Thereafter, when the first cache server 100-1 transmits a data packet including a fragment of content data to the client 300, the first cache server 100-1 counts the number of transmissions for the transmitted fragment. Can be done. For example, if the first fragment is transmitted, the number of transmissions is '1'. If the transmission number information, which has been previously transmitted and counted, is stored in the memory or data fragment storage, it can be added to '1'. will be.

In one embodiment of the present invention, for example, when the number of transmissions for a particular piece exceeds the threshold as a result of counting in the first cache server 100-1, the remaining pieces except for the particular piece are deleted from the memory. .

Then, when the first cache server 100-1 receives a request for content data from any client, the first cache server 100-1 re-determines the number of pieces of data to be included in one data packet, and among the remaining pieces of the first cache server ( 100-1) fragments stored in each partition of the data fragment storage and fragments stored in each partition of the second cache server to n-th cache servers 100-2 to 100-n, respectively, corresponding to the content data. In order for the pieces to be sequentially loaded back into the memory of the first cache server (100-1). In addition, the first cache server 100-1 may include the pieces of pieces that are already present in the internal memory and the pieces that have been reloaded into the payload of the data packet by the number of pieces that have been re-determined and transmitted to an arbitrary client.

In another embodiment of the present invention, for example, among the pieces whose transmission frequency exceeds the threshold, the remaining pieces of the pieces except for the pieces stored in each partition of the first cache server 100-1 data piece storage are stored in the first cache server ( 100-1) Deleted from the internal memory and sending a message to the second cache server to the n-th cache server (100-2 to 100-n), the second cache server to the nth of the pieces of the transmission frequency exceeds the threshold Requests not to delete the pieces retrieved from the cache servers 100-2 to 100-n from the memory inside the second cache servers to the nth cache servers 100-2 to 100-n.

Then, when the first cache server 100-1 receives a request for content data from any client, the first cache server 100-1 re-determines the number of pieces of data to be included in one data packet, and among the remaining pieces of the first cache server ( 100-1) Fragments stored in each partition of the data fragment storage, fragments in the memory of the second cache server to n-th cache server (100-2 to 100-n), and second cache server to n-th cache Among the pieces that do not exist in the memory inside the servers 100-2 to 100-n, pieces stored in each division area of the second cache server to the nth cache server 100-2 to 100-n data fragment storage are stored in the content data. According to the order of each piece corresponding to the sequentially loaded back into the memory of the first cache server (100-1). In addition, the first cache server 100-1 may include the pieces of pieces that are already present in the internal memory and the pieces that have been reloaded into the payload of the data packet by the number of pieces that have been re-determined and transmitted to an arbitrary client.

If the number of transfers between data fragments for the same video content is large, for example, it is because a large number of clients watch the front part of the video and do not watch the back part, so the data fragments corresponding to the front part are relatively large. Can be. In this case, the pieces of data corresponding to the relatively high number of transmissions are left in the memory of the first to n-th cache servers 100-1 to 100-n, and the second to n-th cache servers ( Each of the 100-2 to 100-n may receive a transfer request from the first cache server 100-1 and quickly transfer a fragment existing in the internal memory to the first cache server 100-1.

The configuration of the first cache server 100-1 according to an embodiment of the present invention will be described in more detail with reference to FIG. 2.

2 is a configuration diagram of the first cache server 100-1 shown in FIG. 1 according to an embodiment of the present invention. 1 and 2, the first cache server 100-1 includes an input / output unit 110, a control module 120, a data fragment storage 130, and a memory 125.

The input / output unit 110 has an interface for transmitting and receiving data.

The data fragment storage 130 is a repository of content data fragments stored in the first cache server 100-1. The data fragment storage 130 has a storage space divided into division areas of a corresponding size to store pieces of content data divided into a predetermined size.

The control module 120 is a module for controlling data stored in the data fragment storage 130 and data transmitted and received through the input / output unit 110. In the present embodiment, the control module 120 divides the content data fragments into each division region of the data fragment storage 130 and each division of the second cache server to the nth cache server 100-2 to 100-n data fragment storage. It contains memory 125 for use in retrieving from the area and transferring it to the client 300.

When the control module 120 of the present invention receives the content data through the input / output unit 110, the control module 120 divides the data into pieces of a predetermined size and divides each division area of the data piece storage 130 and the second cache server to the nth cache server 100. -2 to 100-n) Distributed storage in each partition of the data fragment storage. In addition, the start address information of each divided area for storing each piece may be stored in the memory 125.

Thereafter, when the client 300 receives the request for the content data, which is divided and stored from the client 300, the number of pieces of content data to be included in each data packet payload transmitted to the client 300 is determined.

The control module 120 distributes the content data fragments distributed in each division region of the data fragment storage 130 and each division region of the second cache server to the n th cache servers 100-2 to 100-n. It is sequentially loaded into the memory 125. For example, each piece of content data may be loaded into the memory 125 with reference to the start address information of the divided region in which each piece held by the memory 125 is stored.

Although the memory 125 is shown as belonging to the control module in the present embodiment, it may be included in a separate device in the first cache server 100-1, and the process of importing a piece of content data into the memory 125 may be described above. Since the description is the same as that, detailed description will be omitted.

If the data fragments loaded into the memory 125 are the same as the determined number of fragments, the control module 120 includes the corresponding number of data fragments in the payload of the data packet and the client 300 through the input / output unit 110. To send. And the control module 120 counts the number of times of transmission of the pieces sent to the client (300).

In one embodiment of the present invention, if the number of transmissions for a particular piece as a result of counting in the control module 120 exceeds the threshold, the control module 120 stores the remaining pieces except the corresponding piece exceeding the threshold memory 125 Delete from Thereafter, when the control module 120 receives a request for content data from any client through the input / output unit 110, the number of pieces to be included in one data packet payload in consideration of network 400 traffic conditions is determined. Decide again.

The remaining pieces, except for the pieces remaining in the memory 125, are stored in the partition area of the data piece storage 130 and the partition area of the second cache server to the nth cache server 100-2 to 100-n data piece storage. The content data fragment is included in the payload of the data packet by the number of pieces determined again, and is transmitted to the arbitrary client.

In another embodiment of the present invention, if the number of transmissions for a particular piece as a result of counting in the control module 120 exceeds the threshold, the control module 120 is stored in the data piece storage 130 of the pieces exceeding the threshold The remaining pieces other than the pieces are deleted from the memory 125. The second cache server through the n-th cache server 100-2 to 100-n transmits a message and the like, and the second cache server through the n-th cache server 100-2 through 100-n among the pieces exceeding the threshold. Requests not to delete any fragments from internal memory.

Thereafter, when the control module 120 receives a request for content data from any client through the input / output unit 110, the number of pieces to be included in one data packet payload in consideration of network 400 traffic conditions is determined. Decide again.

The remaining fragments other than the fragments remaining in the memory 125 are divided into a partition area of the data fragment storage 130 and a memory and data fragment storage of the second cache servers to the nth cache servers 100-2 to 100-n. The content data pieces are included in the payload of the data packet by the number of pieces determined again, and then transferred to the memory 125 from the area.

A process of transmitting the content data to the client 300 by the control module 120 will be described in more detail with reference to FIG. 3.

3 is a block diagram of a piece information table 310 referenced by the control module 120 shown in FIG. The fragment information table 310 is a control module 120 by the control module 120 from the partition of the data fragment storage 130 and the partition of the second cache server to the n-th cache server (100-2 to 100-n) data fragment storage memory ( 125 is an example of a table referred to to load the content data piece and deletes the content data piece from the memory 125, and the technical scope of the present invention is not limited to the following description. The piece information table 310 may be stored in the data piece storage 310, the memory 125, or a storage inside / outside the first cache server 100-1 according to an embodiment.

Referring to the fragment information table 310 in the present embodiment, the fragment information table 310 includes content, data size, fragment size, fragment information, transmission frequency, and cache server item.

The content item is a field including content information of the content data fragment stored in the data fragment storage 130.

The data size item is a field indicating the size of each content.

The fragment size item is a field indicating a constant size obtained by dividing the content.

The piece information item is a field including data piece information corresponding to each piece of content data.

The transmit count item is a field indicating the number of times each piece has been sent to the client.

The cache server item is an item that represents the cache server where each piece is stored.

Referring to the content item, it can be seen that the content stored in the data fragment storage 130 is 'movie.mpg'. The data size of the 'movie.mpg' content is '10 Mb (Megabyte) '.

Referring to the fragment size item, it can be seen that the 'movie.mpg' content is divided into pieces having a data size of '1 Mb' and stored. Referring to the fragment information item, the fragment in which the 'movie.mpg' content is divided is 'slice1', 'slice2', 'slice3', ..., 'slice10'.

The number of pieces of 'slice1' is' 100 ', the number of pieces of' slice2 'is' 100', the number of pieces of 'slice3' is' 40 ', ..., the number of pieces of' slice10 'is' 20 '. In addition, the 'slice1' piece to the first cache server 100-1, the 'slice2' piece to the second cache server 100-2, the 'slice3' piece to the third cache server, ..., ' It can be seen that the slice10 'slices are stored in the tenth cache server, respectively.

In the fragment information table 310, the 'movie.mpg' content is divided into slices of 'slice1', 'slice2', 'slice3', ..., 'slice10' sequentially from the beginning, and 'slice1' and 'slice2' The number of transfers of the fragments is the largest and the number of transfers is decreasing as the slice10 slices go. For example, if the content of 'movie.mpg' is transmitted to an arbitrary client and the user or the user frequently stops watching the content while watching, the 'corresponding to the front of the' movie.mpg 'content in time The number of transfers of the slice1 'and' slice2 'pieces may be relatively high, and for the pieces with a slow playing order, the number of transfers may be relatively small.

For example, when the first cache server 100-1 transmits 'movie.mpg' content to the client 300, the control module 120 includes two fragments in one data packet payload. Suppose you decide. Thereafter, the control module 120 sequentially loads the 'slice1' and 'slice2' pieces into the memory 125 in each division area of the data piece storage 130.

To this end, the memory 125 or the piece information table 310 may include the start address information of the partition area of the data piece storage 130 in which the 'slice1' piece is stored, and the control module 120 includes the start address. Each piece may be loaded into the memory 125 with reference to the information. At this time, since the 'slice2' fragment is stored in the second cache server 100-2, the first cache server 100-1 requests and receives the 'slice2' fragment from the second cache server 100-2. The partition area storing the 'slice2' fragments, which are loaded into the memory 125 and held by the memory 125 of the first cache server 100-1 or the memory inside the second cache server 100-2. With reference to the start address information of may be loaded into the memory 125.

Then, the 'slice1' and 'slice2' pieces are included in the payload of the data packet and transmitted to the client 300, and the transmission count information '100' corresponding to the 'slice1' and 'slice2' pieces of the fragment information table 310 is displayed. Count the value and store the value of '101'. If the threshold is set to '80', the counting results in the number of transfers of the 'slice1' and 'slice2' pieces exceeding '80', leaving the 'slice1' and 'slice2' pieces in the memory 125.

Thereafter, the control module 120 sequentially loads the pieces from slice3 to slice10 into the memory 125 by two pieces and transmits the pieces to the client 300 in the same manner. If two pieces are transmitted at the same time and the number of transmissions corresponding to each of the two pieces is counted, the pieces are deleted from the memory 125 to free space in the memory 125 unless the threshold '80' is exceeded. do. In this case, the control module 120 may dynamically set the threshold value with reference to the remaining storage capacity of the memory 125.

When the first cache server 100-1 receives a request for 'movie.mpg' content from an arbitrary client, the control module 120 determines the number of pieces to be included in one data packet payload. . If three pieces are determined by the number of pieces, the control module 120 loads pieces of slice3 from the partition of the third cache server data piece store into the memory 125 and remains in the memory 125. The pieces of 'slice1', 'slice2' and 'slice3' imported from the third cache server are included in the payload of the data packet and transmitted to the client 300. Since the process of transmitting the remaining pieces corresponding to 'movie.mpg' is as described above, a detailed description thereof will be omitted.

In another embodiment of the present invention, the first cache server 100-1 is the second cache server 100-2 to the second cache server 100-2, leaving only the slice of 'slice1' in the memory. By requesting not to delete the fragment of 'slice2' stored in the memory, the fragment of 'slice2' may exist in the memory of the second cache server 100-2.

Thereafter, the first cache server 100-1 receives a request for content data of 'movie.mpg' from an arbitrary client, and if the determined number of pieces is 3, the first cache server 100-1 sends a request to the second cache server 100-2. 2 cache server 100-2 requests to send a slice of 'slice2' stored in the internal memory, and requests to send a 'slice3' piece stored in a partition of the third cache server data fragment storage to the third cache server. . When the first cache server 100-1 receives the pieces and loads them into the memory 125, 'slice1' existing in the memory 125 and 'loaded from the memory of the second cache server 100-2' A slice of 'slice3', which is loaded from a slice region of slice2 'and a third cache server data slice store, is included in the payload of the data packet and transmitted to the client 300. Since the process of transmitting the subsequent pieces is as described above, a detailed description thereof will be omitted.

A content transmission process according to an embodiment of the present invention will be described in more detail with reference to FIG. 4.

4 is a flowchart illustrating a content transmission method according to an embodiment of the present invention. 1, 2 and 4, when the first cache server 100-1 receives content data from the content server 200 (S410), the first cache server 100-1 is divided into a predetermined size and divided into a predetermined size. The cache server 100-1 stores the partitioned area of the data fragment storage 130 and the partitioned area of the second cache server to the nth cache servers 100-2 to 100-n data fragment storage (S415).

In step S415, the start address information of the partition area for storing each piece is stored in the memory 125 and / or the second cache server to the nth cache server 100-2 to 100-n in the cache server 100. Can be stored in memory, etc.

As described above, the predetermined size dividing the content data in step S415 may be the size obtained by dividing the maximum transmission unit by a specific integer.

When the first cache server 100-1 receives a request for the content data stored in the step S415 from the client 300 (S420), the first cache server 100-1 includes the data packet payload transmitted to the client 300. The number of pieces of content data is determined (S425). In step S425, for example, the number of fragments may be determined in consideration of network 400 traffic. In addition, as described above, the number of pieces may be determined so that the product of the predetermined size dividing the content data and the number of pieces does not exceed the maximum transmission unit.

In the first cache server 100-1 determining the number of pieces in step S425, the pieces of content data stored in step S415 are sequentially loaded into the memory 125 according to the order of each piece corresponding to the content data. In step S430, the number of pieces determined in step S425 is included in one data packet payload and transmitted to the client 300 (S435). In operation S430, the start address information of each divided area stored in the memory of each cache server may be referred to and loaded into the memory.

In addition, the first cache server 100-1 counts the number of transmissions of the pieces transmitted to the client 300 in step S435 (S440), and counts the remaining pieces except the pieces in which the counted number of transmissions exceeds the threshold. Is deleted from the memory 125 (S445).

Then, when receiving a request for the content data stored in step S415 from any client (S450), as in step S425, the number of pieces to be included in the data packet payload to be sent to any client is determined again. (S455).

In operation S455, the first cache server 100-1 determining the number of fragments may delete the fragments deleted in operation S445 and the partition area of the data fragment storage 130 of the first cache server 100-1 and the second portion. From the partition of the data fragment storage of the cache server to n-th cache server (100-2 to 100-n) to the memory 125 inside the first cache server (100-1) (S460), memory 125 In consideration of the order of the fragments remaining in the previous step, the number of fragments determined in step S455 is included in the payload of the data packet and transmitted to an arbitrary client (S465).

In another embodiment of the present invention, in step S445, the data fragment storage 130 of the first cache server 100-1 is present among the pieces whose transmission counts exceeded the threshold as counted in step S440. The remaining pieces except for the piece may be deleted from the memory 125. In this case, by transmitting a message to the second cache server to the n-th cache server (100-2 to 100-n) it may be requested not to delete a fragment whose transmission frequency exceeds the threshold in the memory of each cache server.

Then, when receiving a request for the content data stored in step S415 from any client (S450), as in step S425, the number of pieces to be included in the data packet payload to be sent to any client is determined again. (S455).

The first cache server 100-1 determining the number of fragments in operation S455 may include a segmentation area of the data fragment storage 130 of the first cache server 100-1 with the fragments deleted in operation S445. Corresponding to the content data from the memory of the second cache server to the n-th cache server (100-2 to 100-n) and the partition area of the second cache server to n-th cache server (100-2 to 100-n) data fragment storage The payload of the data packet is sequentially loaded into the memory 125 according to the order of the pieces (S460) and the number of pieces determined in step S455 in consideration of the order of the pieces remaining in the memory 125. Included in the transmission to any client (S465).

According to the present invention, the cache server loads a predetermined number of pieces of content stored in a plurality of cache server data stores into a memory and transmits them to the client, thereby minimizing fragmentation loss caused when the content transmission is incompletely terminated and network traffic. Depending on the situation, it is possible to select and transfer the appropriate amount of data.It is possible to store the divided data fragments in the divided partitions and to load and transfer them in memory quickly. Content data can be transferred at a faster rate than loaded into memory and can be delivered to the client quickly by storing a relatively high volume of fragments in memory.

1 is a block diagram of a content delivery system according to an embodiment of the present invention.

FIG. 2 is a configuration diagram of the first cache server shown in FIG. 1.

3 is a block diagram of a piece information table referenced by the control module shown in FIG.

4 is a flowchart illustrating a content transmission method according to an embodiment of the present invention.

Claims (22)

The cache server receives the content data from the content server, divides the data into pieces of a predetermined size, and distributes the data to each partition of the cache server data fragment storage and the partitions of one or more other cache server data fragment storage. A storing step of storing; Determining, by the cache server, the number of pieces to be included in one data packet when receiving a request for the content data from a client; And The cache server sequentially stores the fragments stored in each partition of the cache server data fragment storage and the fragments stored in each partition of the other cache server data fragment storage in the order of each fragment corresponding to the content data. And transmitting the received fragments to the client by including the retrieved fragments in the payload of the data packet by the determined number of fragments. Content transmission method using a plurality of cache server comprising a. The method of claim 1, wherein the storing step And storing the start address information of each divided area for storing the divided pieces in the corresponding memory. The method of claim 2, wherein the transmitting step And the stored fragments are loaded into the memory by referring to the start address information of each division area. The method of claim 1, wherein the predetermined size Content transfer method using a plurality of cache server, characterized in that the maximum transfer unit (maximum transfer unit) divided by a specific integer size. The method of claim 1, wherein the determining step Determining the number of fragments in consideration of network traffic. The method of claim 1, wherein the determining step And determining the number of fragments such that the number of fragments multiplied by the predetermined size does not exceed a maximum transmission unit. The method of claim 1, wherein the transmitting step The cache server makes a request to the other cache server for a piece stored by each of the other cache servers, and each of the other cache servers receiving the request stores pieces stored in each partition of the corresponding data piece storage. And a plurality of cache servers, wherein the plurality of cache servers are loaded into an included memory and transmitted to the cache server, and the cache server receives a fragment transmitted from the other cache server and loads the fragment into the memory inside the cache server. Content transmission method using. The method of claim 7, wherein the transmitting step And counting the number of transmissions of the transmitted fragments at the same time as the transmission. The method of claim 8, wherein the transmitting step And deleting the remaining fragments from the memory except for the fragments whose transmission frequency exceeds the threshold as a result of the counting. The method of claim 9, wherein the cache server After deleting the remaining pieces from the memory and receiving a request for the content data from any client, the number of pieces to be included in one data packet is determined again, and the cache server data piece storage of the remaining pieces is deleted. The fragments stored in each partition of the partition and each of the other cache server data fragment storage of the fragments in the order of each of the pieces corresponding to the content data are sequentially loaded back into the memory of the cache server, Using a plurality of cache server, characterized in that for each of the number of pieces of the re-determined pieces, the payload of the data packet including the fragments existing in the memory of the cache server and the remaining fragments to be recalled to the arbitrary client. How to send content. The method of claim 8, wherein the transmitting step Among the fragments whose transmission count exceeds the threshold, the remaining fragments, except the fragments stored in each division area of the cache server data fragment storage, are deleted from the memory inside the cache server, and a message is sent to the other cache server. And requesting not to delete, from the memory inside the other cache server, the piece loaded from the other cache server among the pieces whose transmission frequency exceeds a threshold. The method of claim 11, wherein the cache server After deleting the remaining pieces from the memory and making the request, when receiving a request for the content data from any client, the number of pieces to be included in one data packet is determined again, and the cache among the remaining pieces is deleted. A fragment stored in each partition of the server data fragment storage, a fragment existing in the memory inside the other cache server, and a fragment not present in the memory inside the other cache server, in each partition of the other cache server data fragment storage. The stored fragments are sequentially reloaded into the memory of the cache server according to the order of the fragments corresponding to the content data, and at the same time, the pieces of the determined fragments are present in the payload of the data packet in the memory of the cache server. Piece and the rest of the pieces Also the content delivery method using a plurality of cache servers, characterized in that transmitting by the arbitrary client. A content server providing content; And When the content data is received from the content server, the content data is divided into pieces of a predetermined size and distributed and stored in each partition of the internal data fragment storage partitioned by the predetermined size and in each partition of the data fragment storage of one or more other cache servers. Upon receiving a request for the content data from a client, the number of fragments to be included in one data packet is determined, and the fragments stored in each partition of the cache server data fragment storage and each partition of the other cache server data fragment storage. The pieces stored in are sequentially loaded into the internal memory according to the order of the pieces corresponding to the content data, and the determined number of pieces are transmitted to the client including the retrieved pieces in the payload of the data packet. A cache server; Content delivery system using a plurality of cache server comprising a. Input and output unit for transmitting and receiving data; A data fragment storage for storing pieces of content data having a predetermined size in a divided area divided by the predetermined size; And When the content data is received through the input / output unit, the content data is divided into pieces of the predetermined size and distributed and stored in each division area of the internal data fragment storage and each division area of one or more other cache server data fragment storages, and from the client through the input / output unit. When the request for the content data is received, the number of pieces to be included in one data packet is determined, and pieces stored in each partition of the data fragment storage and pieces stored in each partition of the other cache server data fragment storage are determined. At the same time as the pieces of the pieces corresponding to the content data are sequentially loaded into the internal memory, at the same time, including the loaded pieces in the payload of the data packet by the number of pieces determined to be transmitted to the client through the input and output unit A control module; Cache server comprising a. The method of claim 14, wherein the control module And storing the start address information of each divided area for storing the divided pieces in the corresponding memory. The method of claim 15, wherein the control module The cache server, characterized in that for loading the stored pieces into the memory with reference to the start address information of each of the divided areas. The method of claim 14, wherein the control module A request is made to each of the other cache servers through the input / output unit for the pieces stored in each of the other cache servers, and each of the other cache servers receiving the request includes pieces stored in each partition of the data fragment storage. And a memory server, which is loaded into a memory and simultaneously transmitted to the input / output unit, wherein the control module receives a fragment transmitted from the other cache server through the input / output unit and loads the fragment into the memory inside the control module. The method of claim 17, wherein the control module And a count of the number of transmissions for the transmitted fragments simultaneously with the transmission. The method of claim 18, wherein the control module And discarding the remaining fragments from the memory except for the fragment whose transmission count exceeds the threshold. The method of claim 19, wherein the control module After deleting the remaining pieces from the memory and receiving a request for the content data from any client through the input / output unit, the number of pieces to be included in one data packet is determined again. Pieces stored in each section of the fragment storage and pieces stored in each section of the other cache server data fragment storage are sequentially loaded back into the memory inside the control module according to the order of each piece corresponding to the content data. At the same time, the pieces of the determined pieces are transmitted to the arbitrary client through the input / output unit including the pieces existing in the memory inside the control module and the remaining pieces loaded in the payload of the data packet. Cache server. The method of claim 18, wherein the control module Among the pieces whose transmission count exceeds the threshold value, the remaining pieces except for the pieces stored in each division area of the data piece storage are deleted from the memory inside the control module, and the message is transmitted to the other cache server to transmit the pieces. And a plurality of cache servers requesting not to delete from the memory inside the other cache server among the pieces whose number of times exceeds the threshold. The method of claim 21, wherein the control module Delete the remaining fragments from the memory and send a message to the other cache server to request that the fragments retrieved from the other cache server from the fragments whose transmission frequency exceeds a threshold not be deleted from the memory inside the other cache server. After receiving a request for the content data from any client through the input / output unit, the number of pieces to be included in one data packet is determined again, and the pieces of the remaining pieces are stored in each divided area of the data piece storage. A piece stored in each division area of the other cache server data piece storage among pieces, pieces existing in the memory inside the other cache server, and pieces not existing in the memory inside the other cache server, respectively; Recall sequentially according to the order of the pieces Simultaneously reloading into the memory inside the control module, the payload of the data packet by the number of the pieces determined again includes the pieces present in the memory inside the control module and the remaining pieces reloaded through the input / output unit. Cache server, characterized in that for transmitting to any client.

Images