KR101907786B1 - Distributed storage method and apparatus for cooperation between multiple helper nodes with cached memory - Google Patents

Abstract

The present invention relates to a distributed storage method for enhancing the effects experienced by a user through inter-node collaboration in the wireless communication of multiple help nodes using cache memory and a device thereof. The method, which is for storing files to multiple help modes included in a cluster in a distributed manner, includes the steps of: determining the length of the entire maximum distance separable (MDS) codes coded to have at least a redundant part among multiple MDS codes while using specific files to be divided into the MDS codes based on an MDS coding scheme as a target; dividing the specific file into the MDS codes based on the entire length of the determined MDS codes; and storing the divided MDS codes in the cache corresponding to the help codes in a distributed manner.

Description

TECHNICAL FIELD [0001] The present invention relates to a distributed storage method and apparatus for enhancing a user's haptic effect through cooperation between nodes in a plurality of auxiliary node wireless communications using a cache memory. [0002] DISTRIBUTED STORAGE METHOD AND APPARATUS FOR COOPERATION BETWEEN MULTIPLE HELPER NODES WITH CACHED MEMORY [

The present invention relates to a technique for distributing files to various auxiliary nodes belonging to a cluster in order to provide data, information, files, etc. related to a service to at least one user terminal belonging to a cluster. More specifically, a plurality of helper nodes having a cache and a user equipment (UE) receiving a service from each of the demodulation nodes (for example, receiving data, files, etc.) And more particularly to cache memory distributed storage techniques that provide optimal performance (e.g., increase in transfer capacity) through cooperation between two or more auxiliary nodes when defined as a subject (cluster).

When a cluster is regarded as one communication entity, one cluster may include one or more help nodes and a receiving node, that is, a user terminal, to which files, data, and the like are provided from an auxiliary node. Here, the auxiliary node may be a relay device such as a gateway rather than a base station, and may be used to perform an auxiliary function for relieving the burden of the base station providing data to the receiving node, and the base station to which the auxiliary node belongs May be represented as a transmitting node.

For example, one cluster may have a certain radius, and may include a plurality of auxiliary nodes having a cache within a certain area. There may be one base station (i.e., a transmitting node) to which a plurality of auxiliary nodes belong. Then, the user terminal belonging to the cluster can receive data from the auxiliary node located near or near the user terminal without receiving the desired data from the base station. That is, there is a problem that the transmission signal of the other cluster acts as an interference to the user, thereby reducing the communication quality.

Also, since the capacity of the cache installed for each auxiliary node is limited, there is a problem of determining which data, files, etc. are to be stored in the limited cache. For example, if a file or data having a high probability of requesting a user terminal from a cache is stored in advance, the user terminal can quickly receive a desired service (i.e., file or data) without having to access a base station. In the existing distributed storage technique, when there are a plurality of files, priority is given to the important files in order of popularity, and the files are sequentially distributed in the cache in order of priority. There is a distributed storage method based on Maximum Distance Separable (MDS) coding with other distributed storage techniques.

The distributed storage method based on the MDS coding is a method in which only concatenated packets of a file are cut out and distributed and stored through MDS coding.

For example, if the original file is composed of k packets, the distributed storage technique based on MDS coding is a distributed storage method capable of recovering the original even if only k packets among n packets after n packets are created Coding technique. In case of MDS coding, it is necessary to save several fragmented file fragments in several auxiliary nodes to maximize the storage efficiency. However, in order to satisfy this requirement, n must be very large. When n is large, The calculation is very complicated and it is difficult to apply to the actual wireless communication environment.

Therefore, in a wireless communication environment in which two or more clusters exist, there is a need for a technology for distributing and storing data (or files, etc.) in a cache appropriately in consideration of characteristics of a wireless communication environment, i.e., a wireless channel.

Korean Patent Laid-Open No. 10-2010-00048130 relates to a cluster-based distributed storage system and a method of operating the same, and discloses a technique of multicasting data into a cluster after distributing data in a cache.

The present invention relates to a technique for distributing files in a cache of an auxiliary node based on MDS coding in a cluster to which a plurality of auxiliary nodes and a user terminal (i.e., a single user) belong.

In addition, based on the Maximum Distance Separable (MDS) coding, files are distributed and stored, and the redundancy of the MDS coding (i.e., the MDS coded MDS code) (E.g., average transmission capacity) according to the technique of the present invention.

A method for distributedly storing files in a plurality of help nodes belonging to a cluster, the method comprising the steps of: extracting a plurality of MDS codes based on MDS (Maximum Distance Separable) Dividing a particular file into a plurality of MDS codes based on the total length of the determined MDS code, determining a total length of the plurality of MDS codes, And distributing the MDS codes to a cache corresponding to each of the plurality of auxiliary nodes.

According to an aspect of the present invention, the dividing into the plurality of MDS codes may divide the specific file into a plurality of MDS codes having the same size.

According to another aspect, the size corresponding to each of the plurality of divided MDS codes of the specific file may be determined based on the number of files to be stored in the cache and the memory size of the cache.

According to another aspect, the step of determining the total length of the MDS code may determine the length based on a size corresponding to each of the plurality of MDS codes into which the specific file is divided.

According to another aspect of the present invention, the step of determining the total length of the MDS code includes: calculating a transmission capacity based on the size corresponding to the MDS code and the length; And determining the length and the size corresponding to the highest transmission capacity among the transmission capacity.

According to another aspect, data corresponding to a redundancy between the plurality of divided MDS codes may be provided to a user terminal requesting a corresponding file based on cooperative communication between the plurality of auxiliary nodes.

According to another aspect of the present invention, data corresponding to a remaining region of the plurality of MDS codes excluding the redundancy between the plurality of MDS codes is allocated to one of the plurality of auxiliary nodes, To the user terminal.

According to another aspect of the present invention, the step of dividing the MDS code into a plurality of files is a step of dividing the file into a plurality of files in order of a predetermined number of files sequentially determined in order of preference with respect to the file having the highest preference among the plurality of files, It can be divided into MDS code.

According to another aspect, the cluster may comprise a plurality of auxiliary nodes with a cache and a single user terminal.

A data distribution storage device for distributedly storing files in a plurality of help nodes belonging to a cluster, the data distribution storage device comprising: means for searching a specific file to be divided into a plurality of MDS codes based on MDS (Maximum Distance Separable) A length determining unit for determining an overall length of a coded MDS code in which at least a part of the plurality of MDS codes is redundant, an MDS code for dividing a specific file into a plurality of MDS codes based on the determined total length of the MDS codes, And a storage control unit for distributing and storing the divided MDS codes into a cache corresponding to each of the plurality of auxiliary nodes.

According to an aspect of the present invention, the MDS code division unit may divide the specific file into a plurality of MDS codes having the same size.

According to another aspect, the size corresponding to each of the plurality of divided MDS codes of the specific file may be determined based on the number of files to be stored in the cache and the memory size of the cache.

According to another aspect, the length determination unit may determine the length based on a size corresponding to each of the plurality of MDS codes into which the specific file is divided.

According to another aspect of the present invention, the length determiner calculates a transmission capacity based on the size corresponding to the MDS code and the length, and calculates a transmission capacity corresponding to the highest transmission capacity among the plurality of transmission capacities calculated by adjusting the length The length and size can be determined.

According to another aspect, data corresponding to a redundancy between the plurality of divided MDS codes may be provided to a user terminal requesting a corresponding file based on cooperative communication between the plurality of auxiliary nodes.

According to another aspect of the present invention, data corresponding to a remaining region of the plurality of MDS codes excluding the redundancy between the plurality of MDS codes is allocated to one of the plurality of auxiliary nodes, To the user terminal.

According to another aspect of the present invention, the MDS code division unit divides the file into a plurality of MDS codes for a predetermined number of files sequentially determined in order of preference with respect to the file having the highest preference among the plurality of files Can be divided.

According to embodiments of the present invention, in distributing and storing files in the cache of the secondary node based on MDS coding in a cluster to which a plurality of secondary nodes and a single user terminal (i.e., a single user) belong, By allowing redundancy, or overlap, of the MDS code, it is possible to effectively determine the MDS code to be stored in the cache in the finite MDS code candidate.

Also, in a distributed storage of files based on Maximum Distance Separable (MDS) coding, the redundancy of the MDS coding (i.e., the MDS coded MDS code) is adjusted so that the MDS code That is, by dividing and storing file fragments, it is possible to improve the hitting probability and communication performance (for example, average transmission capacity) in the cluster.

1 is a diagram illustrating a cluster including a plurality of auxiliary nodes in an embodiment of the present invention.
FIG. 2 is a flowchart illustrating a data distribution storage relationship between a base station, an auxiliary node, and a user terminal according to an embodiment of the present invention.
FIG. 3 is a diagram illustrating the configuration of data stored in a cache memory corresponding to each of auxiliary nodes in an embodiment of the present invention.
FIG. 4 is a flowchart illustrating a data distribution storage method according to an embodiment of the present invention.
5 is a block diagram illustrating a configuration of a data distribution storage device in an embodiment of the present invention.
6 is a diagram illustrating a structure of a cache memory between auxiliary nodes that provide a file using an MRT, according to an embodiment of the present invention.
FIG. 7 is a graph showing a transmission capacity increase rate when a memory size is small, in an embodiment of the present invention.
FIG. 8 is a graph showing a transmission capacity increase rate when a memory size is large, in an embodiment of the present invention. FIG.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings.

In a wireless communication environment in which two or more auxiliary nodes and a single user terminal form a communication cluster, a technique for improving the hitting probability and communication performance (for example, average transmission capacity) . In particular, the present invention uses MDS (Maximum Distance Separable) coding to distribute at least one file among a plurality of files stored in a base station to a plurality of auxiliary nodes corresponding to a specific cluster among clusters belonging to a base station And a technique for distributing and storing divided MDS codes (i.e., file fragments / packets) in a cache of each of a plurality of auxiliary nodes by adjusting the redundancy of the MDS coding.

In the present embodiments, the transmitting node represents a base station, and the auxiliary node is a relay device such as a gateway existing between a base station and a user equipment. The relaying device may be a relay device for relieving the burden of providing data to the receiving node Can be used to perform auxiliary functions. For example, the transmitting node may be represented as a data distributed storage device. The receiving node may indicate a user terminal receiving a packet (i.e., a file fragment) corresponding to the request file from at least one of the transmitting node (i.e., the base station) to which the corresponding cluster belongs and the auxiliary nodes belonging to the cluster. For example, the receiving node may include a smart phone, a tablet PC, a notebook, and the like.

In addition, the auxiliary node has a cache memory, and stores frequently requested files from a base station corresponding to a cluster to which the auxiliary node belongs, that is, a user equipment (UE) . For example, the files may include data files stored in a general file system, multimedia data such as video and music, control data for connection of a communication system, and the like.

In the present embodiments, the hitting probability indicates the probability that a file requested by the user terminal exists in the cache of the auxiliary node, that is, the probability that a portion of the file requested by the user terminal is stored in the cache .

In these embodiments, the MDS code may represent the packet as a fragmented packet (i.e., a piece of file) based on MDS coding to distribute the file stored in the base station to the cache of each of the plurality of auxiliary nodes. The size of the MDS code to be stored in the cache of each of the auxiliary nodes can be represented by S, and the redundancy of at least a part of the areas corresponding to the MDS codes divided for the specific file is L < / RTI > For example, when file 1 is divided into two, the redundancy between MDS code 1 and MDS code 2 can be represented by L. [ An MDS code corresponding to a portion of the file, that is, the entire length of the MDS coding, can be stored in each cache.

In these embodiments, the size of the MDS code to be stored in the cache of each of the auxiliary nodes may indicate the size of the memory for the MDS code to be stored in each cache. That is, the partitioning may be performed in the MDS coding based on the size of the entire cache size that can store the MDS code. For example, the MDS codes (i.e., packets) corresponding to S are randomly stored in each cache. At this time, since S codes of all the files are stored in each cache, they can overlap and be stored. That is, some of the MDS codes stored in each cache among the entire lengths of the MDS-coded files may be stored redundantly.

In the present embodiments, it is assumed that the auxiliary node has a cache memory. It is assumed that the minimum value of L indicating the total length of the MDS-coded MDS code by allowing redundancy is 1, and that there is no finite maximum value.

1 is a diagram illustrating a cluster including a plurality of auxiliary nodes in an embodiment of the present invention.

1, the cluster may include a plurality of auxiliary nodes 111 and 112 and a user terminal 113. In FIG. It is assumed that each of the transmission nodes 111, 112, 121, and 122 has one antenna.

The packets related to the file requested by the user terminal 113 belonging to the cluster 100 (for example, the MDS code corresponding to the file fragment) are dispersed in the cache of each of the auxiliary node 1 111 and the auxiliary node 2 112 May be stored. Then, the user terminal 113 can receive the file directly through the auxiliary node 1 (111) and the auxiliary node 2 (112) without transmitting the file from the base station. At this time, different packets (i.e., MDS codes) may be stored in the cache of each of the auxiliary node 1 111 and the auxiliary node 2 112, and at least one of the entire areas corresponding to each of the MDS codes Some of the areas may be duplicated and stored in each cache. That is, a part of the MDS code stored in the cache of each of the auxiliary node 1 111 and the auxiliary node 2 112 may be redundantly stored. Then, the user terminal 113 can simultaneously receive codes corresponding to the redundancy stored in the auxiliary nodes 111 and 112 based on the cooperative communication, and codes corresponding to the non-overlapping areas And can be provided from the auxiliary node storing the code.

FIG. 2 is a flowchart illustrating a data distribution storage relationship between a base station, an auxiliary node, and a user terminal according to an embodiment of the present invention.

In step 210, the base station 201, which is a transmitting node, transmits a specific file to a plurality of MDS codes (that is, a plurality of MDS codes) based on MDS coding so that at least a part of the specific files among the plurality of files, , File fragment).

In step 220, the base station 201 may control the plurality of divided MDS codes to be distributedly stored in the cache of the plurality of auxiliary nodes. For example, a plurality of MDS codes may be broadcast to a plurality of auxiliary nodes.

In step 230, each of the auxiliary nodes 202 may store the MDS code provided from the base station 201 in its own cache.

If the user terminal 230 wants to receive a specific file (that is, to receive a specific service such as music or movie) in step 240, the auxiliary node 202 transmits a transmission request for a specific file to the user terminal 230 As shown in Fig.

In step 250, since the MDS code which is a part of the requested file is stored in advance in the cache of the auxiliary node 202, the auxiliary node 202 does not separately request the MDS code corresponding to the requested file to the base station 201 To the user terminal (203).

의 선호도(예컨대, 인기도)

는 지프 분포(Zipf distribution)를 따르며, 피크 상수(peakness constant)는 Zipf exponent인

에 따라 아래의 수학식 1과 같이 결정될 수 있다.As described above, the files can be divided and stored in the cache of each of the auxiliary nodes. Then, the number of files stored in each cache can be represented as F _c, in order to improve the capacity, that is, the length L of the MDS code which is stored the number of file F _c and a cache store in order to increase the transmission capacity Lt; / RTI > At this time, the minimum value of L is 1, and a finite maximum value may not exist. If there are a total of F files, each file

(E.g., popularity)

Is followed by a Zipf distribution and the peakness constant is a Zipf exponent

1 < / RTI >

[Equation 1]

개의 파일의 일부분을 저장할 수 있다. 예를 들어, 복수개의 파일들 중 선호도가 가장 높은 파일을 기준으로 선호도가 상대적으로 높은 순서에 따라 순차적으로 결정된 일정 개수(즉, F_c)의 파일의 일부분(MDS 코딩된 코드 중 일부에 해당하는 데이터, 즉, 복수개로 분할된 MDS 코드들 중 적어도 하나의 MDS 코드)가 캐시 메모리에 저장될 수 있다. 여기서, F_c개의 파일들 각각의 일부 영역이 저장된 보조 노드 각각의 캐시 메모리 상태는 아래의 도 3을 참고하여 후술하기로 한다. Each of the auxiliary nodes may have one transmitting antenna, and each of the auxiliary nodes may have a high preference in the cache

A portion of the file can be stored. For example, a portion of a file of a predetermined number (i.e., F _c ) sequentially determined in order of preference relative to a file having the highest preference among a plurality of files (a part of the MDS- Data, i.e., at least one MDS code among a plurality of divided MDS codes) may be stored in the cache memory. Here, the cache memory state of each of the auxiliary nodes, in which a partial area of each of the F _c files is stored, will be described later with reference to FIG.

FIG. 3 is a diagram illustrating the configuration of data stored in a cache memory corresponding to each of auxiliary nodes in an embodiment of the present invention.

FIG. 3 illustrates a state of a cache memory in which MDS codes (i.e., packets) corresponding to a randomly selected file fragment are stored for two auxiliary nodes. For example, in FIG. 3, files 1 and 2, which are stored in the cache of each of the two auxiliary nodes (H _1,1 , H _1,2 ) that provide the request file to the user terminal 113 of FIG. 1, .

(311)과

(313)의 크기가 동일할 수 있다. 보조 노드 1(H_1,1)의 캐시는 파일 1의 일부에 해당하는 MDS 코드(

, 330)와 파일 2의 일부에 해당하는 MDS 코드(

, 350)를 S(0<S<1)의 양만큼 저장하고 있을 수 있다. 마찬가지로, 보조 노드 2(H_1,2)의 캐시는 파일 1의 일부에 해당하는 MDS 코드(

, 340)와 과 파일 2의 일부에 해당하는 MDS 코드(

, 360)를 S(0<S<1)의 양만큼 저장하고 있을 수 있다. 이에 따라, 이에 따라, MDS 코드(, 330)와 MDS 코드For example, when MDS is coded to distribute the file 1 (File 1, 310) in a plurality of caches, only a part of the entire MDS-coded codes can be separately stored in the cache of the plurality of auxiliary nodes. At this time, the size of the MDS code stored in each cache (for example, the size of the packet) may be the same. In other words, the MDS code can be divided so that file 1 310 is an equal size in MDS coding. For example, when the divided MDS code 1

(311) and

(313) may be the same size. The cache of the auxiliary node 1 (H _1,1 ) is an MDS code corresponding to a part of the file 1

, 330) and an MDS code corresponding to a part of the file 2

, 350) by S (0 < S < 1). Similarly, the cache of the auxiliary node 2 (H _1,2 ) stores the MDS code corresponding to a part of the file 1

, 340) and an MDS code corresponding to a part of the file 2

, 360) by S (0 < S < 1). Accordingly, the MDS code ( , 330) and the MDS code

(311)+

(312)-중복되는 영역(313)을 나타낼 수 있다.The cache of BonozoD 1 storing the MDS code 330 and the cache of the auxiliary node 2 storing the MDS code 350 are arranged such that at least a part of the divided MDS codes are overlapped The same data corresponding to the area 313 may be stored. At this time, the user terminal can request the transfer of the file 1 310. The sum of the unique portion of the MDS code 330 constituting the file 1 and the unique portion of the MDS code 340 constituting the file 1 is the size of the original data A file requested by the user terminal can be provided using only the auxiliary node 1 and the auxiliary node 2 storing the file 1 310 (i.e., the hitting probability can be predetermined as 1 in FIG. 3) . Here, the sum of the unique portions of the MDS code 330 constituting the file 1 and the MDS code 340 constituting the file 1 is

(311) +

(312) -overlapping region (313).

The auxiliary nodes belonging to one cluster can sequentially transmit their MDS codes to the user terminal at predetermined regular time intervals. For example, each of the auxiliary nodes may transmit data corresponding to a unique portion different from the MDS code that has been transmitted sequentially in a time division manner at a predetermined time interval to the user terminal. In this case, the overlapping area (i.e., overlapping area) between the MDS codes partitioned for the same file corresponds to a section in which all the auxiliary nodes belonging to the same cluster know the same information, so that multiple-input single-output (MISO) All of the auxiliary nodes can cooperatively transmit the corresponding file. The operation of controlling the data corresponding to the overlapped area to be transmitted to the user terminal based on the cooperative communication will be described later with reference to FIG.

FIG. 4 is a flowchart illustrating a data distribution storage method in an embodiment of the present invention, and FIG. 5 is a block diagram illustrating a configuration of a data distribution storage device in an embodiment of the present invention.

4 and 5 illustrate the case where the data distributed storage device is a base station, but this corresponds to the embodiment, and the data distribution storage device may be any one of the plurality of auxiliary nodes. For example, any one of the auxiliary nodes storing the original data requested by the user terminal may operate as a data distributing apparatus.

5, the data distribution storage apparatus 500 may include a length determination unit 510, an MDS code division unit 520, and a storage control unit 530. Each of the plurality of auxiliary nodes belonging to the data distributed storage device 500 may have a cache, which is a memory, and may have one transmission antenna. 4 may be performed by the length determination unit 510, the MDS code division unit 520, and the storage control unit 530, which are components of FIG.

In step 410, the length determining unit 510 determines the length of the coded symbol length of the MDS code allowed to be redundant between the MDS codes, based on the MDS coding, on a specific file to be divided into a plurality of MDS codes, L can be determined. That is, when performing the MDS coding on a specific file, the length determiner 510 can determine the total length L of the MDS-coded code.

In step 411, the length determining unit 510 may determine the size S of the MDS code to be divided, with respect to a specific file. At this time, the length determining unit 510 can determine the size S of the MDS code based on the number of files to be stored in the cache and the size of the cache memory.

For example, the length determining unit 510 may calculate the size S of a plurality of MDS codes while changing the number of files to be stored in the cache. At this time, the length determining unit 510 may divide the MDS codes into equal sizes for the same file. For example, when the file 1 is divided into N, the sizes of the N MDS codes may be the same.

In step 412, when the size S of the MDS code is determined, the length determination unit 510 may calculate the total length L of the MDS code optimized for the size S. [ For example, the length L optimized for each of the plurality of sizes S can be calculated. In other words, as the size S is changed, the length L of the overlapping area can also be adjusted. For example, the Sdp-optimized L may be differentiated with respect to L by a formula related to the transfer rate (sum rate) derived based on the following Equations 2 and 3, and the L value in which the differentiated formula is 0 is optimized L, respectively.

In step 413, the length determining unit 510 may calculate the transmission capacity based on the calculated size S and the total length L of the calculated MDS code. At this time, as the length L is adjusted, a corresponding plurality of transmission capacities can be calculated. For example, if M size S is calculated and L optimized (i.e., length adjusted) for each of M size S is calculated, M transmission capacities can be calculated.

In step 414, the length determiner 510 may determine the length L of the redundant area corresponding to the highest transmission capacity among the plurality of transmission capacities and the size S of the MDS code as L and S to be finally applied to the MDS coding .

For example, if F = 10 and M = 3, there may be cases where Fc = 4, 5, 6 and S = 0.75, 0.6, 0.5. . At this time, the optimized L value can be calculated for each case of S = 0.75, 0.6, 0.5 and the highest sum rate (for example, transmission rate) of the sum rate Capacity) can be selected.

In operation 420, the MDS code division unit 520 may divide the specific file into a plurality of MDS codes based on the finally determined length L and the size S of the MDS code.

In step 430, the storage control unit 530 may control the divided MDS codes to be distributedly stored in the cache corresponding to each of the plurality of auxiliary nodes. For example, the storage control unit 530 may broadcast the divided MDS codes to the auxiliary nodes so that the MDS codes are distributedly stored in the cache of each of the auxiliary nodes.

6 is a diagram illustrating a structure of a cache memory between auxiliary nodes that provide a file using an MRT, according to an embodiment of the present invention.

6, the length determining unit 510 determines the size S (602, 603) of the MDS code for dividing the cache memory size of the auxiliary node (for example, a possible size of the cache memory 601) ). For example, the plurality of auxiliary nodes may each have a cache of the same memory size, and the data distributed storage device 500, which is a base station, may know the size of the cache memory in advance. The length determiner 510 may determine an overall length L (604) that allows redundancy between MDS codes for a specific file. In other words, if S is 1 and is divided into 10, then L may not be 10 and may be less than 10 due to redundancy. At this time, the length L may be adjusted depending on whether or not the redundancy is allowed. For example, the more the redundancy 605 is allowed, the shorter the length L 604, and the shorter the redundancy 605 is, the longer the length L 604 may be.

In FIG. 6, since the overlapping area 605 corresponds to an area storing the same information in the cache of all the auxiliary nodes belonging to the same cluster, all the auxiliary nodes can cooperate and transmit the corresponding file to the user terminal. At this time, the cooperation can be performed based on MRT (Maximum Likelihood Transmission), and the transmission capacity obtainable through collaboration can be expressed by Equation (2) below.

&Quot; (2) &quot;

는신호대 잡음비(SNR)를 나타낼 수 있다. 보조 노드의 개수가 3개 이상인 경우, 보조 노드의 개수에 해당하는 h가 수학식 2에 추가될 수 있다. 중복 영역(605)에 해당하는 데이터는 보조 노드 1 및 보조 노드 2를 통해 사용자 단말로 동시에 전송될 수 있다. 이처럼, 동시에 데이터를 전송하는 방식을 MRT 모드로 표현할 수 있다. 다시 말해, 수학식 2는 MRT 모드로 데이터 전송 시의 전송 용량을 나타낼 수 있다.In equation (2) the data to secondary nodes is the two are co corresponds to the overlapping region 605 as representing the transmission capacity of the case of providing to the user terminal, h ₁ is the channel information of the secondaries 1, h ₂ are two secondaries The channel information can be displayed. And,

SNR &lt; / RTI &gt; If the number of auxiliary nodes is three or more, h corresponding to the number of auxiliary nodes may be added to Equation (2). Data corresponding to the redundant area 605 can be simultaneously transmitted to the user terminal through the auxiliary node 1 and the auxiliary node 2. As described above, the method of transmitting data at the same time can be expressed by the MRT mode. In other words, Equation (2) can represent the transmission capacity at the time of data transmission in the MRT mode.

At this time, since data corresponding to the remaining area excluding the overlapping area 605 in the MDS-coded code related to the file requested by the user terminal can not be transmitted simultaneously, each auxiliary node can transmit to the user terminal in the TDMA scheme. This data transmission scheme can be expressed in a single mode. The transmission capacity in the case of transmitting data to the user terminal in the single mode can be expressed by Equation (3) below.

&Quot; (3) &quot;

는신호대 잡음비(SNR)를 나타낼 수 있다. In Equation (3), h _i denotes channel information of the i-th auxiliary node,

SNR &lt; / RTI &gt;

As described above, at the present time when a file stored in the cache of each of the auxiliary nodes (that is, MDS code divided by MDS coding) is transmitted to the user terminal, data of overlapping areas between MDS codes is transmitted, The MRT mode and the single mode can be distinguished according to whether the auxiliary nodes transmit the data of the auxiliary nodes, and the data corresponding to the overlapping areas of the auxiliary nodes simultaneously are provided to the user terminal based on cooperation, The total average transmission capacity by providing data corresponding to the non-overlapping area to the user terminal can be expressed by Equation (4) below.

&Quot; (4) &quot;

는 MRT 모드로 전송하는 시간 부분(time portion),

는 싱글(single) 모드로 전송하는 시간 부분(time portion)을 나타낼 수 있다. 그리고,

는, MRT 모드로 전송 시의 평균 전송 용량,

는 싱글 모드로 전송 시 얻을 수 있는 평균 전송 용량을 나타낼 수 있다.In Equation (4), Pr [hit] is a file existence probability (hitting probability) inside the cluster,

A time portion to transmit in MRT mode,

May represent a time portion of transmission in a single mode. And,

The average transmission capacity at the time of transmission in the MRT mode,

Can represent the average transmission capacity that can be obtained when transmitting in the single mode.

In Equation (4), the file existence probability Pr [hit] can be expressed by the total length L of the MDS coded data and the size S of the MDS code as shown in Equation (5) below.

&Quot; (5) &quot;

은 아래의 수학식 6과 같이 표현될 수 있다.Then, according to the relationship between the total length L and the size S of the MDS code,

Can be expressed by Equation (6) below.

&Quot; (6) &quot;

에 기초하여

와

는 아래의 수학식 7과 같이 표현될 수 있다.Then, the signal-to-noise ratio (SNR)

On the basis of

Wow

Can be expressed by Equation (7) below.

&Quot; (7) &quot;

At this time, if the size of cache memory for each auxiliary node is M, the number of storable files can be changed from M + 1 to 2M max. Accordingly, the size S of the MDS code can be determined according to the number of files and the size of the cache memory. For example, the size S of the MDS code can be calculated for each of the cases where the number of files is M + 1, M + 2, ..., 2M. Then, the optimized L is determined for each calculated size S, and S and L corresponding to the maximum transmission capacity among the transmission capacities calculated based on S and L can be determined as the final S and L for MDS coding a specific file .

FIG. 7 is a graph showing a transmission capacity increase rate when the memory size is small, in an embodiment of the present invention. FIG.

에 따른 기존 기법 대비 전송 용량 증가율을 나타낼 수 있다. 예컨대, 도 6에서, 캐시 메모리의 크기M은 3인 경우(즉, M=3)를 가정한다.In FIG. 7, it is assumed that the first half and the second half of each file corresponding to the length of 0.5 without using the MDS code are distributed to the respective auxiliary nodes with respect to the size M of the cache memory (conventional technique) The transmission capacity increase rate when the MDS coded data is distributedly stored based on S and L can be shown. That is, SNR and Zipf exponent,

The rate of increase of the transmission capacity compared to the conventional technique. For example, in FIG. 6, it is assumed that the size M of the cache memory is 3 (i.e., M = 3).

If the total number of files is F, a maximum min (2M, F) number can be stored. The user terminal can receive a file fragment (that is, a divided MDS code) from each secondary node and restore the file requested by the terminal have. At this time, the obtainable transmission capacity can be expressed by Equation (8) below.

&Quot; (8) &quot;

Table 1 below shows the performance compared with each parameter, that is, the degree of improvement in transmission capacity.

Parameter value Number of helper / user 2 Number of user One Number of file 10 Signal to noise ratio ( SNR ) -10 to -30 dB Zipf dist . exponent 0 to 2 Channel model Rayleigh  flat fading

와

간의 차이가 감소함을 확인할 수 있다. 그리고, 지프 지수

가 증가할수록 다수의 파일 저장에 의한 성능 이득이 감소하므로, 낮은 SNR과 높은 지프 지수에 대한 성능(예컨대, 평균 전송 용량)이 증가함을 확인할 수 있다. 이와 반대되는 구간, 즉, SNR이 높고 지프 지수

가 낮을수록 상대적으로 성능(예컨대, 평균 전송 용량)이 감소함을 확인할 수 있다.Referring again to FIG. 7, when the cache memory size M = 3 (that is, when the cache memory size is small), as the SNR increases

Wow

The difference between them decreases. Then, the Jeep index

It can be seen that the performance (for example, the average transmission capacity) for the low SNR and the high Jeep index increases. In contrast, the SNR is high and the Jeep index

(For example, the average transmission capacity) is decreased.

FIG. 8 is a graph showing a transmission capacity increase rate when a memory size is large, in an embodiment of the present invention. FIG.

In FIG. 8, the transfer capacity increase rate in the case of the size M = 7 of the cache memory can be shown.

Referring to FIG. 8, it can be seen that the performance (that is, the transfer capacity) increases when the memory size is relatively larger (M = 7) than when the cache memory size M = 3 as shown in FIG.

=2인 경우, 도 7과 같이 M=3일 때의 성능(전송 용량)은 약 40%이고, 도 8와 같이 M=7일 때의 성능(전송 용량)은 약 100%로 성능 향상이 훨씬 높은 점을 확인할 수 있다. 이에 따라, 기존 기법 대비 약 2.1배의 전송 용량 확보가 가능할 수 있다.In other words, since the maximum S value increases as the memory size increases, it can be seen that a relatively high performance improvement occurs when M = 3. For example, SNR = 0, Jeep index

= 2, the performance (transmission capacity) when M = 3 is about 40% as shown in FIG. 7, and the performance (transmission capacity) when M = 7 is about 100% High points can be confirmed. Accordingly, it is possible to secure a transmission capacity of about 2.1 times that of the conventional technique.

As described above, in the cluster environment composed of the plurality of auxiliary nodes having the cache memory and the user terminal, it is possible to adjust the overall length (i.e., generation length) of the MDS code to achieve the optimum performance have. By using a wireless communication technique (for example, transmission in the MRT mode) that allows redundancy between MDS codes to obtain additional performance, a portion of the finite MDS codes to be stored in the auxiliary node can be determined, And can be applied to an environment such as a small cell or a femtocell used in a network.

The method according to an embodiment may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions to be recorded on the medium may be those specially designed and configured for the embodiments or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention is not limited to the disclosed exemplary embodiments. For example, it is to be understood that the techniques described may be performed in a different order than the described methods, and / or that components of the described systems, structures, devices, circuits, Lt; / RTI &gt; or equivalents, even if it is replaced or replaced.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (17)

A method for distributedly storing files in a plurality of help nodes belonging to a cluster, which is performed by a data distribution storage device including a length determination unit, an MDS code division unit, and a storage control unit,
Wherein the length determining unit determines the length of the entire MDS code that is coded by redundancy so that at least some of the plurality of MDS codes are overlapped on a specific file to be divided into a plurality of MDS codes based on Maximum Distance Separable (MDS) Determining a length;
Dividing a specific file into a plurality of MDS codes based on the total length of the determined MDS code in the MDS code division unit; And
Distributing a plurality of divided MDS codes in a cache provided in each of the plurality of auxiliary nodes in the storage control unit
/ RTI &gt; The method according to claim 1,
Wherein the dividing into the plurality of MDS codes comprises:
And dividing the specific file into a plurality of MDS codes having the same size. The method according to claim 1,
The size corresponding to each of the plurality of MDS codes into which the specific file is divided is determined based on the number of files to be stored in the cache and the memory size of the cache
And storing the data. The method according to claim 1,
Wherein determining the total length of the MDS code comprises:
Determining the length based on a size corresponding to each of the plurality of MDS codes into which the specific file is divided
And storing the data. 5. The method of claim 4,
Wherein determining the total length of the MDS code comprises:
Calculating a transmission capacity based on the size corresponding to the MDS code and the length; And
Determining the length and size corresponding to the highest transmission capacity among the plurality of transmission capacities calculated by adjusting the length,
/ RTI &gt; The method according to claim 1,
Wherein data corresponding to a redundancy between the plurality of divided MDS codes is provided to a user terminal requesting a corresponding file based on cooperation communication between the plurality of auxiliary nodes. The method according to claim 1,
Data corresponding to a remaining area excluding the redundancy between the plurality of MDS codes is provided to the user terminal from any one of the plurality of auxiliary nodes with respect to the plurality of divided MDS codes, And storing the data. The method according to claim 1,
The step of dividing into the MDS code comprises:
Dividing a file into a plurality of MDS codes for a predetermined number of files sequentially determined in order of preference based on a file having the highest preference among the plurality of files
And storing the data. The method according to claim 1,
Wherein the cluster comprises a plurality of auxiliary nodes with a cache and a single user terminal. CLAIMS What is claimed is: 1. A data distribution storage apparatus for distributedly storing files in a plurality of help nodes belonging to a cluster,
Determining a length of determining the total length of a redundantly coded MDS code for a specific file to be divided into a plurality of MDS codes based on MDS (Maximum Distance Separable) coding, at least a part of the plurality of MDS codes being redundant part;
An MDS code division unit for dividing a specific file into a plurality of MDS codes based on the total length of the determined MDS code; And
A storage controller for distributively storing a plurality of divided MDS codes in a cache provided in each of the plurality of auxiliary nodes,
Lt; / RTI &gt; 11. The method of claim 10,
The MDS code division unit,
And dividing the specific file into a plurality of MDS codes having the same size. 11. The method of claim 10,
The size corresponding to each of the plurality of MDS codes into which the specific file is divided is determined based on the number of files to be stored in the cache and the memory size of the cache
And the data distribution storage device. 11. The method of claim 10,
The length determining unit may determine,
Determining the length based on a size corresponding to each of the plurality of MDS codes into which the specific file is divided
And the data distribution storage device. 14. The method of claim 13,
The length determining unit may determine,
Calculating a transmission capacity based on the size corresponding to the MDS code and the length, and determining the length and size corresponding to the highest transmission capacity among the plurality of transmission capacity calculated by adjusting the length
And the data distribution storage device. 11. The method of claim 10,
Wherein data corresponding to a redundancy between the plurality of divided MDS codes is provided to a user terminal requesting a corresponding file based on cooperative communication between the plurality of auxiliary nodes. 11. The method of claim 10,
Data corresponding to a remaining area excluding the redundancy between the plurality of MDS codes is provided to the user terminal from any one of the plurality of auxiliary nodes with respect to the plurality of divided MDS codes, And the data is stored in the storage unit. 11. The method of claim 10,
The MDS code division unit,
Dividing a file into a plurality of MDS codes for a predetermined number of files sequentially determined in order of preference based on a file having the highest preference among the plurality of files
And the data distribution storage device.

Images