KR101973537B1 - Apparatus and method for managing of heartbeat exchange period in the hadoop distributed file system - Google Patents

Abstract

The present invention relates to a method for managing a heartbeat exchange period in a Hadoop distribution file system. The method for managing a heartbeat exchange period comprises the steps of: (a) receiving the number of worker nodes in a cluster of the system and a network bandwidth of the system in a server device; (b) calculating and updating a heartbeat exchange period in the server device by using the number of the received worker nodes and the network bandwidth; and, (c) determining whether to transmit information on the updated heartbeat exchange period to the worker node in the cluster in consideration of a difference between the updated heartbeat exchange period and the heartbeat exchange period before the update is conducted in the server device.

Description

[0001] APPARATUS AND METHOD FOR MANAGING HEARTBEAT EXCHANGE PERIOD IN THE HADOP DISTRIBUTED FILE SYSTEM [0002]

The present invention relates to an apparatus and method for managing a heartbeat exchange period in a Hadoop distributed file system.

With the development of the information industry, the development of technologies for storing and analyzing big data, which is a vast amount of data, is being actively developed. Distributed processing technology is a key technology in handling such big data.

In a large-scale distributed system, it is difficult for an individual node to know the status of the entire system and the network usage status. Therefore, the central server determines and manages the exchange cycle of the heartbeat, which is an information exchange message, The individual nodes transmit heartbeats to the central server according to the determined period at the central server.

1 is a diagram schematically illustrating a flow of a heartbeat exchange transmission in a conventional Hadoop Distributed File System (HDFS).

Referring to FIG. 1, in a conventional Hadoop distributed file system, which is a distributed processing framework for large amount of data, a server device as a central server manages a heartbeat exchange period at a fixed cycle, so that worker nodes, Can be transmitted to the server apparatus every fixed period (a constant cycle). In this way, each worker node (data node) can notify the central server (name node) of the heartbeat and provide its own state change information along with the notification of its existence.

However, in the conventional Hadoop distributed file system, since each worker node transmits a heartbeat to a central server at fixed intervals, even if a state change occurs in a worker node after reading or writing is performed to each worker node block, Can not transmit its changed status information to the central server as soon as the status change occurs, and can transmit its changed status information after the maximum period of the set fixed period has passed. This affects the file I / O execution time of the Hadoop distributed file system, which causes the performance degradation of the entire system.

The background technology of the present application is disclosed in Korean Patent Laid-Open Publication No. 10-2015-0082521.

SUMMARY OF THE INVENTION The present invention has been made to solve the above problems of the conventional art and it is an object of the present invention to provide a Hadoop distributed file system, And an object of the present invention is to provide an apparatus and method for managing a heartbeat exchange period in a Hadoop distributed file system capable of solving a problem causing degradation.

It is to be understood, however, that the technical scope of the embodiments of the present invention is not limited to the above-described technical problems, and other technical problems may exist.

According to an aspect of the present invention, there is provided a method of managing a heartbeat exchange period in a Hadoop Distributed File System according to an embodiment of the present invention includes the steps of: (a) Receiving a number of worker nodes in the cluster and a network bandwidth of the system; (b) calculating and updating a heartbeat exchange period in the server device using the number of the received worker nodes and the network bandwidth; And (c) receiving, at the server device, information on the updated heartbeat exchange period and the updated heartbeat exchange period in consideration of the difference between the updated heartbeat exchange period and the heartbeat exchange period before the update is performed, And determining whether to transmit to the mobile station.

In addition, the step (c) may determine whether to transmit to the worker node considering whether the difference between the updated heartbeat exchange period and the heartbeat exchange period of the previous time is equal to or greater than a preset change amount.

In addition, the steps (a) to (c) may be repeated periodically.

In addition, (c) may determine to transmit information on the updated heartbeat exchange period to the worker node when it is determined that the difference is equal to or greater than a preset change amount.

In the step (b), the heartbeat exchange period may be calculated by further considering an average heartbeat size of a heartbeat transmitted from a worker node in the cluster.

이고, 여기서, TP _HB 는 하트비트 교환 주기,

는 계수, N _D 는 클러스터 내 워커노드의 수, PS _HB 는 평균 하트비트 크기, B _n 은 네트워크 대역폭을 나타낸다.In the step (b), the heartbeat replacement period may be calculated as a minimum value satisfying Equation (1)

, Where TP _HB is the heartbeat swap period,

N _D is the number of worker nodes in the cluster, PS _HB is the average heartbeat size, and B _n is the network bandwidth.

Also, the coefficient may be set in consideration of a CRC / Packer header while converting bytes into bits per second (bps).

Also, the coefficient may be 10.

Meanwhile, an apparatus for managing a heartbeat exchange period in a Hadoop Distributed File System according to an exemplary embodiment of the present invention includes an input unit for receiving a number of worker nodes in a cluster and a network bandwidth of the system; A calculating unit for calculating and updating a heartbeat exchange period using the number of the received worker nodes and the network bandwidth; And a determination to determine whether to transmit information on the updated heartbeat exchange cycle to the intracluster worker node in consideration of the difference between the updated heartbeat exchange period and the heartbeat exchange period before the update is performed Section.

The determining unit may determine whether to transmit to the worker node considering whether the difference between the updated heartbeat replacement period and the heartbeat replacement period of the previous time is equal to or greater than a preset change amount.

In addition, the number of worker nodes and the input of the network bandwidth in the input unit, the calculation of the heartbeat exchange period in the calculating unit, and the determination of whether to transmit to the worker node in the determination unit can be periodically repeated.

In addition, the determining unit may determine to transmit information on the updated heartbeat exchange period to the worker node when the difference is determined to be equal to or greater than a preset change amount.

In addition, the calculating unit may calculate the heartbeat exchange period by further considering an average heartbeat size of a heartbeat transmitted from a worker node in the cluster.

이고, 여기서, TP _HB 는 하트비트 교환 주기,

는 계수, N _D 는 클러스터 내 워커노드의 수, PS _HB 는 평균 하트비트 크기, B _n 은 네트워크 대역폭을 나타낸다.Also, the calculating unit may calculate the heartbeat swap period with a minimum value satisfying the following equation (2), and the equation

, Where TP _HB is the heartbeat swap period,

N _D is the number of worker nodes in the cluster, PS _HB is the average heartbeat size, and B _n is the network bandwidth.

In addition, the coefficient may be set in consideration of a CRC / Packer header while converting bytes into bits per second (bps).

Also, the coefficient may be 10.

The above-described task solution is merely exemplary and should not be construed as limiting the present disclosure. In addition to the exemplary embodiments described above, there may be additional embodiments in the drawings and the detailed description of the invention.

According to the above-described task resolution means of the present invention, the heartbeat exchange period is calculated and updated using the number of worker nodes in the cluster and the network bandwidth, and the updated heartbeat exchange period and the heartbeat exchange period It is possible to minimize the response speed of the entire system and effectively improve the system performance by determining whether to transmit information on the updated heartbeat exchange period to the worker node.

According to the above-described task solution, the updated heartbeat replacement period can maximize the performance of an application that requires file I / O.

According to the above-described task solution of the present invention, the updated heartbeat exchange period can optimize the information exchange period (i.e., the heartbeat exchange period) between the server apparatus and the worker nodes, thereby performing the same operation It can be done with minimum time.

However, the effects obtainable here are not limited to the effects as described above, and other effects may exist.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a diagram schematically illustrating a flow of a heartbeat exchange transmission in a conventional Hadoop distributed file system. FIG.
FIG. 2 is a diagram schematically illustrating the structure of an entire system including a heartbeat exchange period management apparatus in the Hadoop distributed file system according to an embodiment of the present invention. Referring to FIG.
3 is a flowchart illustrating a method for managing a heartbeat exchange period in the Hadoop distributed file system according to an embodiment of the present invention.

Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings so that those skilled in the art can easily carry out the present invention. It should be understood, however, that the present invention may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. In the drawings, the same reference numbers are used throughout the specification to refer to the same or like parts.

Throughout this specification, when an element is referred to as being "connected" to another element, it is intended to be understood that it is not only "directly connected" but also "electrically connected" or "indirectly connected" "Is included.

It will be appreciated that throughout the specification it will be understood that when a member is located on another member "top", "top", "under", "bottom" But also the case where there is another member between the two members as well as the case where they are in contact with each other.

Throughout this specification, when an element is referred to as " including " an element, it is understood that the element may include other elements as well, without departing from the other elements unless specifically stated otherwise.

This paper describes a Hadoop distributed file system that improves the performance of the overall system by optimizing the heartbeat period of information exchange messages of the central server and other nodes (i.e., worker nodes) in a big data distributed processing platform such as Hadoop. And more particularly, to an apparatus and method for managing a heartbeat exchange period in a heartbeat exchange period.

FIG. 2 is a diagram schematically illustrating the structure of an entire system including a heartbeat exchange period management apparatus in the Hadoop distributed file system according to an embodiment of the present invention. Referring to FIG.

Referring to FIG. 2, the heartbeat replacement period management apparatus 10 in the Hadoop Distributed File System (HDFS) according to an exemplary embodiment of the present invention may be a server, A central node, a central server, a central resource management server, a namenode, or the like. The server device 10 may receive the heartbeat from the worker nodes 20 in the cluster of the Hadoop distributed file system. Here, the worker node 20 may be otherwise referred to as a data node (Datanode).

A basic explanation of the heartbeat is as follows. In the distributed processing system, since work is divided, it is necessary to share the progress and the result of work. At this time, there is no need to share data every time during the distribution processing. The reason for this is that there are no major changes to the worker node and it takes time for the result to be completed. Accordingly, a heartbeat, which is an information exchange message for sharing data of a certain period, is used in a distributed processing system. In the distributed processing system, the heartbeat periodically exchanges information between nodes or processes, thereby enabling efficient data processing.

The Hadoop Distributed File System (HDFS) to which the heartbeat replacement period management apparatus 10 according to an embodiment of the present invention is applied is a file system supporting a distributed environment. The server apparatus 10 and the worker node (20).

The server device 10 may refer to a server that manages metadata of a file system. The worker node 20 stores actual data blocks and supports reading and writing of data when a client needs it.

For example, in the HDFS, when a client attempts to read or write a specific file, it can access the server device 10 to acquire the position of the block corresponding to the specific file (IP address of the worker node). Then, the worker node corresponding to the position of the obtained block can be connected to perform reading and writing of the required block.

Here, because of the scalability problem between the server apparatus 10 and the worker node 20, the server apparatus 10 does not send a message to the worker node 20 first, but transmits the message to the worker node 20 And may transmit a message in response to a heartbeat, which is a message transmitted from the worker node 20, if necessary. Communication between the server device 10 and the worker node 20 can be performed. In this case, as described above, in the conventional HDFS, since the worker node can transmit the heartbeat in a fixed period, there is a problem that it affects the file I / O execution time of the HDFS and causes the performance degradation of the whole system.

Accordingly, in order to solve the conventional problem, the present invention provides a method and apparatus for dynamically checking (checking) an element that affects the determination of a replacement cycle of a heartbeat in an HDFS, calculating an optimal heartbeat cycle, The present invention proposes a heartbeat exchange period management apparatus 10 that enables transmission of a heartbeat from a node. A detailed description of the server device 10 which is the heartbeat-exchange-period management device 10 is as follows.

The server device 10 may include an input unit 11, a calculation unit 12, and a determination unit 13. [

The input unit 11 can receive the number of the worker nodes in the cluster of the Hadoop distributed file system and the network bandwidth.

That is, the server device 10 can dynamically check the number of worker nodes and the network bandwidth for calculation of the optimal heartbeat replacement period based on Equation (1) to be described later, and the input unit 11, And information on the network bandwidth.

The calculating unit 12 may calculate and update the heartbeat exchange period using the number of worker nodes received from the input unit 11 and the network bandwidth. In other words, the calculating unit 12 can calculate the optimized heartbeat exchange cycle (optimal heartbeat exchange cycle) by using the number of worker nodes and the network bandwidth, and the previously set heartbeat exchange cycle (Which may mean the heartbeat exchange period of time before the current heartbeat is exchanged) with the calculated optimum heartbeat exchange period.

Further, the calculating unit 12 may calculate the heartbeat exchange period by further considering the average heartbeat size of the heartbeat transmitted from the worker node in the cluster. In other words, the calculating unit 12 can calculate the heartbeat exchange period in consideration of the number of worker nodes in the cluster, the network bandwidth, and the average heartbeat size.

The calculating unit 12 can calculate the heartbeat replacement period with the minimum value satisfying the following expression (1). At this time, the minimum value of the heartbeat exchange period satisfying Equation (1) may be referred to as an optimal heartbeat exchange period.

[Equation 1]

는 계수를 나타낸다. N _D 는 하둡 분산 파일 시스템의 클러스터 내 워커노드의 수를 나타낸다. PS _HB 는 평균 하트비트 크기(Average heartbeat size)를 나타내며, 단위는 일예로 바이트(Bytes)일 수 있다. B _n 은 네트워크 대역폭을 나타내며, 단위는 bps일 수 있다. 이때, 계수는 바이트(Bytes)를 초당 비트수(bit-per-second, bps)로 변환하면서 CRC/Packer header를 고려하여 설정될 수 있다. 일예로, 계수는 10으로 설정될 수 있다.Where TP _HB represents the heartbeat swap period, and the unit may be seconds, for example.

Represents a coefficient. N _D represents the number of worker nodes in the cluster of the Hadoop distributed file system. PS _HB represents the average heartbeat size, and the unit may be, for example, Bytes. B _n represents the network bandwidth, and the unit may be bps. At this time, the coefficient may be set in consideration of the CRC / Packer header while converting the Bytes into bits per second (bps). For example, the coefficient may be set to 10.

Specifically, the average transmission and processing time of the average heartbeat, that is, the average transmission and processing time of one heartbeat, may satisfy Equation (2).

&Quot; (2) "

, PS _HB 및 B _n 에 대한 설명은 앞서 설명했으므로, 이하 중복되는 설명은 생략하기로 한다.Where T _HB represents an average heartbeat transmission and processing time, and the unit may be seconds, for example. Also,

, PS _HB, and B _n have been described above, so that the following description will be omitted.

In addition, since the number ( N _HB ) of heartbeats that the server apparatus 10 can process per second is the reciprocal of the average heartbeat transmission and processing time T _HB , N _HB is defined to satisfy the following equation (3) .

&Quot; (3) "

In addition, the number of times the server device 10 receives heartbeats per second from the worker nodes may be defined as N _D / TP _HB .

According to this, N _HB must satisfy the following expression (4).

&Quot; (4) "

Here, Equation (4) can be expressed differently from Equation (5), and Equation (1) can be derived from Equation (5).

&Quot; (5) "

That is, the approximate loads of the server apparatus 10 determined according to the heartbeat exchange cycle and to what extent the heartbeat exchange cycle can be reduced (in other words, in calculating the optimum heartbeat exchange cycle, The limit setting for how much the replacement period can be shortened to a certain extent) can be defined through Equations (1) to (5).

The decision unit 13 determines whether or not the updated heartbeat exchange cycle information (i.e., the value of the updated heartbeat exchange cycle) in consideration of the difference between the updated heartbeat exchange cycle and the heartbeat exchange cycle before the update is performed, To the worker node in the cluster.

The determination unit 13 determines whether or not the information about the updated heartbeat exchange cycle in consideration of whether or not the difference between the updated heartbeat exchange cycle and the heartbeat exchange cycle at the previous time point is equal to or greater than a predetermined change amount (e.g., k% To the worker node (20).

In other words, the decision unit 13 determines whether or not the difference between the updated heartbeat-replacement cycle and the heartbeat-replacement cycle at the previous time has occurred at a large change by a preset variation amount, To the worker node (20). Here, the predetermined change amount k may be set in units of percent (%) as an example, but is not limited thereto.

When it is determined that the difference between the updated heartbeat exchange cycle and the heartbeat exchange cycle at the previous time is equal to or greater than a predetermined change amount, the determination unit 13 transmits information about the updated heartbeat exchange cycle to the worker node 20 You can decide. Based on this determination, the server device 10 may send information about the updated heartbeat swap period to the worker node 20.

At this time, in consideration of the scalability problem between the server apparatus 10 and the worker node 20, the server apparatus 10 updates the updated heartbeat exchange cycle in response to the heartbeat transmitted from the worker node 20 Can be transmitted. In other words, when it is determined to send the information about the updated heartbeat exchange period to the worker node 20 in the decision unit 13, the server apparatus 10 receives the heartbeat from the worker node 20, A response message to a heartbeat (i.e., a message indicating that a heartbeat has been received) may be transmitted to the worker node 20. At this time, information about the updated heartbeat exchange period may be included in the response message and transmitted. In addition, the transmission of the information about the updated heartbeat exchange period from the server apparatus 10 to the worker node can be continuously performed until the server apparatus 10 newly calculates the optimum heartbeat exchange period.

The worker node 20 having received the updated information on the heartbeat exchange period from the server device 10 can transmit the heartbeat to the server device 10 with the updated heartbeat exchange period. That is, the worker node 20 which has received the information on the updated heartbeat exchange period re-updates the value of the heartbeat exchange period to the value of the updated heartbeat exchange period, The heartbeat can be transmitted to the server apparatus 10 according to the period. In other words, when the worker node 20 receives the information about the updated heartbeat exchange period from the server device 10, the worker node 20 transmits the heartbeat to the server device 10 with the optimized heartbeat exchange period based on the updated heartbeat exchange period 10).

On the other hand, the input of the number of worker nodes and the network bandwidth in the input unit 10, the calculation and update of the heartbeat exchange period in the calculation unit 12, and the determination of whether or not to transfer to the worker node in the decision unit 13, And may be periodically repeated in units of a preset cycle. In other words, each process performed in the input unit 10, the calculation unit 12, and the determination unit 13 so that the optimum heartbeat replacement cycle can be applied (set) to the Hadoop distributed file system is periodically . ≪ / RTI > Here, the unit of the predetermined period may mean a unit of several tens minutes, several hours, and the like, but the present invention is not limited to this, and various periods can be set.

In other words, the calculation process of the heartbeat-replacement cycle through the calculation unit 12 is periodically performed in units of a preset cycle, and thus the calculated heartbeat-replacement cycle can be updated. At this time, , The server device 10 can newly receive the information on the network bandwidth and the number of worker nodes and calculate the heartbeat exchange period.

The present application is based on the assumption that the server device 10 calculates and updates an optimized heartbeat exchange cycle (i.e., a message exchange cycle) considering an environment such as network bandwidth, the number of worker nodes in a cluster, The response speed in the Hadoop distributed file system can be minimized by transmitting information on the updated (or computed) heartbeat exchange period to the worker node when the calculated change amount of the heartbeat exchange period is greater than a preset change amount .

Also, the present invention can calculate the minimum value of the heartbeat exchange period between nodes (i.e., between the server apparatus and the worker node) within a range that does not adversely affect the network performance (i. E. It is possible to minimize the response speed of the entire system.

We can dynamically determine the network situation and the number of worker nodes belonging to the cluster, and calculate the optimized heartbeat exchange cycle based on this. From this, we can maximize the performance of applications that require file I / O. In addition, the present invention can determine an optimized heartbeat exchange period in the server device 10 other than the worker node, and can effectively improve the performance of the application through adjustment to the optimum heartbeat exchange period.

In other words, the Hadoop distributed file system is a model in which a central server manages worker nodes that operate on hundreds of machines, but in many cases it is also used for the operation of small clusters. In the case of such a small cluster, the exchange period of the heartbeat should be shorter than the basic setting value, thereby effectively reducing the task execution time. Therefore, we can optimize the heartbeat exchange interval considering the network bandwidth and the number of worker nodes (machines) in the cluster.

This aspect of the present invention can perform the same job with the minimum amount of time in the same hardware as the prior art, by optimizing the information exchange period (i.e., the heartbeat exchange period) between the server apparatus 10 and the worker nodes 20. [ Accordingly, the present invention can be effectively applied to an operation requiring a real-time response (for example, a database query). That is, the present invention can be applied to real-time systems. Also, the present invention can be applied to Hive which is a system that enables Hadoop's HDFS to be handled using a query similar to SQL.

Hereinafter, the operation flow of the present invention will be briefly described based on the details described above.

3 is a flowchart illustrating a method for managing a heartbeat exchange period in the Hadoop distributed file system according to an embodiment of the present invention.

The method of managing the heartbeat exchange period shown in FIG. 3 may be performed by the heartbeat exchange period management apparatus (i.e., server apparatus 10) described above. Therefore, even if omitted below, the description of the method of managing the heartbeat exchange period can be similarly applied to the description of the heartbeat exchange period management apparatus (i.e., the server apparatus 10).

Referring to FIG. 3, in the method of managing the heartbeat exchange period in the Hadoop distributed file system according to the embodiment of the present invention, the number of the worker nodes and the network bandwidth in the cluster of the system in the server device can be inputted (S11).

Next, in step S12, the server device can calculate and update the heartbeat exchange period using the number of worker nodes and the network bandwidth input in step S11.

In addition, in step S12, the heartbeat replacement period can be calculated by further considering the average heartbeat size of the heartbeats transmitted from the worker node in the cluster.

In addition, in step S12, the heartbeat replacement period can be calculated to the minimum value satisfying Equation (1). In Equation (1), the coefficient may be set considering the CRC / Packer header while converting the byte into bits per second (bps), and the coefficient may be 10. At this time, the description related to Equation (1) has been described in detail in detail above, and a duplicate explanation will be omitted below.

Next, a method for managing a heartbeat exchange period in the Hadoop distributed file system according to an embodiment of the present invention is a method for managing the heartbeat exchange period in the Hadoop distributed file system according to an embodiment of the present invention, And determining whether to send information about the updated heartbeat swap period to the in-cluster worker node.

At this time, in the step of determining whether or not to transmit to the worker node, it may be determined whether to transmit to the worker node considering whether the difference between the updated heartbeat exchange period and the heartbeat exchange period of the previous time is equal to or greater than a preset change amount .

Specifically, in the step of determining whether or not to transmit to the worker node, it is possible to determine whether the updated heartbeat exchange cycle is equal to or greater than a predetermined change amount with respect to the heartbeat exchange cycle of the previous time (S13) Lt; RTI ID = 0.0 > heartbeat < / RTI > exchange period to the worker node.

At this time, when it is determined that the difference between the updated heartbeat replacement cycle and the heartbeat replacement cycle at the previous time is equal to or greater than a predetermined change amount (S13-Y), in the step of determining whether to transmit to the worker node, It may decide to send information about the period to the worker node. Accordingly, in step S14, the server apparatus can transmit information on the updated heartbeat exchange period to the worker node. On the other hand, when it is determined that the difference between the updated heartbeat replacement period and the heartbeat replacement period at the previous time is not equal to or more than the preset change amount (S13-N), the process can return to step S11.

Meanwhile, in the method of managing a heartbeat exchange period in the Hadoop distributed file system according to an embodiment of the present invention, a step S11 of receiving input, calculating and updating a heartbeat exchange cycle S12, The determining step may be repeated periodically.

In the above description, steps S11 to S14 may be further divided into further steps or combined into fewer steps, according to embodiments of the present disclosure. Also, some of the steps may be omitted as necessary, and the order between the steps may be changed.

The method for managing the heartbeat exchange period according to an exemplary embodiment of the present invention 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 recorded on the medium may be those specially designed and constructed for the present invention 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 present invention, and vice versa.

It will be understood by those of ordinary skill in the art that the foregoing description of the embodiments is for illustrative purposes and that those skilled in the art can easily modify the invention without departing from the spirit or essential characteristics thereof. It is therefore to be understood that the above-described embodiments are illustrative in all aspects and not restrictive. For example, each component described as a single entity may be distributed and implemented, and components described as being distributed may also be implemented in a combined form.

The scope of the present invention is defined by the appended claims rather than the detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalents should be construed as being included within the scope of the present invention.

10: Server device
11: Input unit
12:
13:
20: Worker node

Claims (17)

A method for managing a heartbeat exchange period in a Hadoop Distributed File System, the method comprising:
(a) receiving, at a server device, the number of worker nodes in the cluster and the network bandwidth of the system;
(b) calculating and updating a heartbeat exchange period in the server device using the number of the received worker nodes and the network bandwidth; And
(c) in the server apparatus, information on the updated heartbeat exchange period is updated in consideration of the difference between the updated heartbeat exchange period and the heartbeat exchange period before the update is performed, to the worker node in the cluster Determining whether to transmit,
, ≪ / RTI &
The step (b)
Calculating the heartbeat replacement period by further considering an average heartbeat size of a heartbeat transmitted from a worker node in the cluster,
Calculating the heartbeat replacement period with a minimum value satisfying the following equation (1)
[Equation 1]

Where TP _HB is the heartbeat swap period,

Where N _D is the number of worker nodes in the cluster, PS _HB is the average heartbeat size, and B _n is the network bandwidth. The method according to claim 1,
The step (c)
Determining whether to transmit to the worker node considering whether the difference between the updated heartbeat exchange period and the heartbeat exchange period of the previous time is equal to or greater than a preset change amount. The method according to claim 1,
Wherein the steps (a) to (c) are repeated periodically. 3. The method of claim 2,
The step (c)
And determine to send information about the updated heartbeat exchange period to the worker node if the difference is determined to be equal to or greater than a predetermined change amount. delete delete The method according to claim 1,
Wherein the coefficient is set in consideration of a CRC / Packer header while converting bytes into bits per second (bps). 8. The method of claim 7,
Wherein the coefficient is ten. A heartbeat exchange period management apparatus in a Hadoop Distributed File System,
An input unit for receiving the number of worker nodes in the cluster and the network bandwidth of the system;
A calculating unit for calculating and updating a heartbeat exchange period using the number of the received worker nodes and the network bandwidth; And
Determining whether to transmit the updated information on the heartbeat exchange period to the in-cluster worker node in consideration of the difference between the updated heartbeat exchange period and the heartbeat exchange period before the update is performed; ,
, ≪ / RTI &
The calculating unit calculates,
Calculating the heartbeat replacement period by further considering an average heartbeat size of a heartbeat transmitted from a worker node in the cluster,
Calculating the heartbeat replacement period with a minimum value satisfying the following equation (2)
&Quot; (2) "

Where TP _HB is the heartbeat swap period,

Where N _D is the number of worker nodes in the cluster, PS _HB is the average heartbeat size, and B _n is the network bandwidth. 10. The method of claim 9,
Wherein,
And determines whether or not to transmit to the worker node considering whether the difference between the updated heartbeat replacement period and the previous-time heartbeat replacement period is equal to or greater than a preset change amount. 10. The method of claim 9,
Wherein a number of worker nodes in the input unit and an input of a network bandwidth, a calculation of a heartbeat exchange period in the calculating unit, and a determination of whether or not to transmit to a worker node in the determination unit are repeated periodically, Exchange period management device. 11. The method of claim 10,
Wherein,
And determines to transmit information about the updated heartbeat exchange period to the worker node when it is determined that the difference is equal to or greater than a preset change amount. delete delete 10. The method of claim 9,
Wherein the coefficient is set in consideration of a CRC / Packer header while converting bytes into bits per second (bps). 16. The method of claim 15,
Wherein the coefficient is 10. A computer readable recording medium having recorded thereon a program for executing the method of any one of claims 1 to 4, 7, and 8 in a computer.

Images