KR102017422B1 - QoE-Aware Video Storage Power Management Method and System Based on Hot and Cold Data Classification - Google Patents

Abstract

Disclosed are a user-perceived quality-aware video storage power management method based on hot and cold data classification, and a system thereof. A method of managing video storage power based on dynamic adaptive steaming over HTTP (DASH) comprises the steps of: transcoding multimedia data into a plurality of versions corresponding to different bit rates; determining, from the data transcoded into the plurality of versions, hot data to be stored in a hot zone in a video storage which is divided into the hot zone and a cold zone; storing the determined hot data in the hot zone of the video storage; storing, in the cold zone of the video storage, remaining data excluding the hot data among the data transcoded into the plurality of versions; and shutting down the cold zone. According to the present invention, it is possible to reduce power consumption.

Description

QoE-Aware Video Storage Power Management Method and System Based on Hot and Cold Data Classification}

Embodiments of the present invention relate to a technique for managing power consumed when storing and accessing data in storage of a video server based on DASH (Dynamic Adaptive Streaming over HTTP), which is a video streaming technique.

Recent developments in network and system technology have made it possible to provide streaming services to a variety of applications such as user-generated content (UCC) and personal game broadcasts. Streaming services have greatly increased Internet video traffic, and streaming servers need large server structures.

For example, YouTube reports that users watch about 5 billion videos a day. Major streaming companies like Netflix, Hulu, and YouTube use Dynamic Adaptive Streaming (DASH) over HTTP. Dynamic Adaptive Streaming over HTTP (DASH) divides each video into segments that are transcoded into different bitrate versions. This allows each client to choose the version of the bitrate that best suits their network conditions, and accurate prediction of network bandwidth is important for providing high quality-of-experience (QoE).

Selecting a bitrate version lower than the actual bandwidth degrades the video quality itself, while selecting a higher bitrate version causes playback stalls or buffering, which is inconvenient for clients. DASH inherently requires a lot of transcoding work. For example, the Internet provides video quality between 1.5 Mbps and 25 Mbps, and 120 transcoding operations may be required for each video clip.

In addition, a large amount of storage space is required to store the converted version, resulting in high disk power consumption to maintain a large amount of storage. The main result of this high power consumption is heat, which can increase the failure rate of the disk drive. In addition, it is essential to reduce storage power consumption because high energy consumption prevents the server from expanding.

Korean Patent Laid-Open Publication No. 10-2016-0008365 relates to a storage medium, a memory system, and a storage area management method in a memory system. Memory that translates the logical address of the write-requested data into a physical address so that the logical address of the hot data is mapped among the cold or less frequent data update data, and writes the write-requested data to the converted physical address of the memory device. A storage area management method in a system is disclosed.

The present invention relates to a technology for reducing power consumption by selectively streaming data of a specific version of data without streaming service of all versions encoded based on DASH, a video streaming technology.

In the DASH (Dynamic Adaptive Streaming over HTTP) based video storage power management method, transcoding a plurality of versions corresponding to different bit rates for multimedia data, wherein the plurality of Determining hot data to be stored in the hot zone in the video storage divided into a hot zone and a cold zone, for the data transcoded into a version, and determining the determined hot data into the video storage. Storing in the cold zone of the video storage the data remaining in the hot zone of the plurality of versions except for the hot data, and down the cold zone. can do.

According to an aspect of the present invention, the cold zone may be configured to bring down the cold zone so that hot data stored in the hot zone is selectively accessed.

According to another aspect, as data corresponding to a specific version stored in the cold zone is requested, the method further includes streaming data corresponding to a transcoded version at a lower bit rate than data stored in the requested cold zone. It may include.

Video storage power management.

According to another aspect, data corresponding to a transcoded version at a lower bit rate than data stored in the cold zone may represent data stored in the hot zone.

According to another aspect, the determining of the hot data, the original data of the multimedia data corresponding to the highest bit rate of the transcoded to the plurality of versions, and the transcoding at the lowest bit rate The determined data may be determined as the hot data.

According to another aspect, storing in the hot zone may include equally allocating a video segment corresponding to the hot data to a plurality of disks constituting the hot zone.

In the DASH (Dynamic Adaptive Streaming over HTTP) based video storage power management system, a transcoding unit transcoding to a plurality of versions corresponding to different bit rates for multimedia data, A determination unit for determining hot data to be stored in the hot zone in the video storage divided into a hot zone and a cold zone, for the data transcoded into a plurality of versions; A storage control unit for storing the data stored in the hot zone of the video storage and storing the remaining data except the hot data among the data transcoded into the plurality of versions in the cold zone of the video storage, and down the cold zone. It may include a power control unit.

According to one aspect, as the data corresponding to a specific version stored in the cold zone is requested, the transmission control unit for streaming transmission of the data corresponding to the transcoded version at a lower bit rate than the data stored in the requested cold zone It may further include.

According to another aspect, the determining unit may include the original data of the multimedia data corresponding to the highest bit rate among the data transcoded into the plurality of versions, and the data that is transcoded at the lowest bit rate into the hot data. Can be determined.

According to another aspect, the storage controller may equally allocate the video segment corresponding to the hot data to a plurality of disks constituting the hot zone.

The present invention divides a disk array of a server into a cold zone and a hot zone without a streaming service of all versions of data transcoded based on DASH, a video streaming technology, and the cold zone spins. By powering down and selectively streaming a particular version of data stored in the hot zone, power consumption can be reduced. That is, after storing the data corresponding to the version to be streamed in the hot zone, and storing the remaining versions of the data in the cold zone, power consumption can be reduced by spinning down the cold zone.

1 is a diagram illustrating a network structure including a video storage power management system according to an embodiment of the present invention.
2 is a block diagram illustrating an internal configuration of a video storage power management system according to an embodiment of the present invention.
3 is a flowchart illustrating a video storage power management method according to an embodiment of the present invention.
4 is a graph illustrating a relationship between power consumption and QoE of 2 TB and 3 TB in an HVP workload according to an embodiment of the present invention.
FIG. 5 is a graph illustrating QoE and power consumption relations of 2 TB and 3 TB in an LVP workload according to an embodiment of the present invention.

Embodiments of the present invention relate to a technology for managing video storage power in consideration of user perception quality based on hot and cold data classification, and particularly, a hot zone for all versions of data transcoded based on DASH. A technique for classifying data to be stored in a cold zone and data to be stored in a cold zone, spin down of a cold zone, and selectively streaming data stored in a hot zone to reduce power consumption. will be.

In the present embodiments, the "disk" may support multiple power modes. Multi-power mode is the active mode where the head reads or writes data, the seek mode the head is browsing, an idle mode where the disk spins at full speed but does not process requests for data, and the disk does not rotate completely It may contain a standby mode that consumes much less energy than the mode but the data on the disk is inaccessible. In this case, in the standby mode, the disk array may be divided into a cold zone and a hot zone to spin down the cold disk in order to minimize power consumption of the storage. At this time, since the data stored in the cold disk cannot be accessed within a predetermined time, it is important to manage the disk composed of the hot and cold zones. In the following, the data stored in the hot zone and the cold zone is classified and managed. Next, the data of the versions to be stored in the hot zone are equally stored, and the processing for the request when the data of the version stored in the cold zone is requested will be described.

In the present embodiments, "version" may be used to distinguish different bit rates. For example, there may be four versions when transcoding at four different bit rates for original data. That is, data of version 1 corresponding to bit rate 1, version 2 corresponding to bit rate 2, version 3 corresponding to bit rate 3, and version 4 corresponding to bit rate 4 may exist.

In the present embodiments, "hot zone" may refer to a hot disk and "cold zone" may refer to a cold disk.

The embodiments may consider three matters as follows.

1. Which video segments will be stored on the hot disk?

2. Which of the hot disks to store the video segments of the versions determined to be stored on the hot disk?

3. Whether the video segment corresponding to the version stored on the cold disk is processed when a request occurs

For these three considerations: 1) a hot data classification algorithm that determines which segments to store on a hot disk, taking into account video segment popularity and user perceived quality; and 2) video segments that balance the workload load between hot disks. Allocation Algorithm, 3) Disc bandwidth allocation algorithm that determines the serviceable version of each requested video segment in order to maximize overall user perception quality.

1 is a diagram illustrating a network structure including a video storage power management system according to an embodiment of the present invention.

Referring to FIG. 1, the video storage power management system 100 may correspond to a streaming server, and may be connected to at least one client terminal 102, 103, 104 through a wired / wireless network.

The video storage power management system 100 may have a built-in video storage 101, and in the case of a streaming server, since the amount of stored multimedia data is enormous, video storage may be externally stored separately from the video storage power management system 100. May exist. Then, the power management system 100 may be connected to the video storage 101 through a wired network, and a plurality of video storage 101 may exist.

FIG. 2 is a block diagram illustrating an internal configuration of a video storage power management system according to an embodiment of the present invention, and FIG. 3 is a flowchart illustrating a video storage power management method according to an embodiment of the present invention.

The video storage power management system 200 of FIG. 2 represents the video storage power management system 100 of FIG. 1, and the video storage power management systems 100 and 200 may include a transcoding unit 210 and a determination unit 220. , A storage controller 230, a power controller 240, and a transmission controller 250.

In addition, each of steps 310 to 360 of FIG. 3 is performed by the transcoding unit 210, the determination unit 220, the storage control unit 230, and the power, which are components of the video storage power management system 200 of FIG. 2. It may be performed by the control unit 240 and the transmission control unit 250.

In operation 310, the transcoding unit 210 may transcode the multimedia data at different bit rates. Then, data corresponding to a plurality of different versions transcoded at different bit rates may be generated. For example, original data with the highest bit rate, data of version 1 transcoded at a bit rate relatively lower than the highest bit rate, data of version 2 transcoded at a bit rate relatively lower than the data of version 1,?, Version k of data transcoded at the lowest bit rate may be generated.

In operation 320, the determiner 220 may determine hot data to be stored in the hot zone in video storage divided into a hot zone and a cold zone, based on the data transcoded into a plurality of versions. have.

In this case, the determination unit 220 may determine data to be stored in the hot disk (ie, in the hot zone) based on the popularity of the data and the user perception quality. For example, the determination unit 220 may determine the original data and the version of the data transcoded at the lowest bit rate as the hot data to be stored in the hot disk. In addition, at least one data may be determined as hot data based on popularity and user perception quality among versions of data existing between the highest bit rate and the lowest bit rate. Here, the video storage may be composed of a plurality of disks, and at least one of the plurality of disks may be classified as a hot disk and a remaining disk except for the hot disk. The number of hot disks and cold disks may be the same, or the number of hot disks may be different from each other, such as more than the number of cold disks, or more than the number of hot disks.

In operation 330, the storage controller 230 may store the determined hot data in a hot zone of the video storage. In this case, the storage controller 230 may determine which hot disk to store the hot data among a plurality of hot disks in order to maintain a balance between the plurality of hot disks constituting the video storage. The hot data can be stored in. That is, the storage controller 230 may control the video segments corresponding to the hot data to be stored in the plurality of hot disks evenly.

In operation 340, the storage controller 230 may store the remaining data (ie, cold data) except for hot data among the data transcoded into a plurality of versions in a cold disk (ie, a cold zone) of video storage. The storage controller 230 may store and maintain the identifier information of the video segment corresponding to the cold data and the identifier information indicating the position of the cold disk in which the video segment is stored.

In operation 350, the power controller 240 may control the cold disk to spin down. In this case, the power control unit 240 may control the cold disk not to rotate to reduce the power consumed due to the disk rotation.

For example, the power control unit 240 may control the cold disk not to rotate until a request for the cold data stored in the cold disk occurs. Even if the cold data request occurs, the power control unit 240 continues until the command to stop the rotation of the cold disk occurs. You can also control the cold disk not to rotate. As such, since the cold disk is in a spin down state where rotation is stopped, when a specific version of the multimedia data request is generated at the client terminal, only the hot data stored in the hot disk may be selectively accessed to be serviced for streaming. .

In operation 350, as the data corresponding to the specific version stored in the cold disk is requested, the transmission control unit 250 transmits data corresponding to the transcoded version at a relatively lower bit rate to the client terminal than the data stored in the requested cold zone. Streaming can be transmitted.

For example, data from version 2 with the second highest bit rate is requested, but data from version 2 is stored on a cold disk, data from version 1 with the highest bit rate and data from version 4 with the fourth highest bit rate on a hot disk. When the data of the version k having the lowest bit rate is stored, the transmission controller 250 transmits the version 4 data closest to the version 2 of the data having the lower bit rate than the version 2 among the data stored in the hot disk to the client terminal. Can be. That is, the data of the version requested by the client terminal and the version of the data having the smallest bit rate difference may be streamed to the client terminal.

Table 1 below may represent an algorithm for determining hot data.

The transmission control unit 250 may access only a hot disk among the plurality of disks constituting the video storage 101. Accordingly, the determination unit 220 may determine the video segment to be stored in the hot disk, and may determine the video segment to be stored in the hot disk based on the hot data determination algorithm shown in Table 1 above. Before explaining the hot data determination algorithm shown in Fig. 1, the concept of the structure of video storage and the parameters used for hot data determination will be described first.

May represent an access probability of video version j of segment i. So for every version of every segment Sum of It can be calculated based on. Popularity of segment i Is It can be calculated as Here, the popularity may indicate how many times the segment i is accessed. For example, more access may be relatively more popular than video segments that access less.

Is the total arrival rate (i.e. arrival rate) for your system. The arrival rate of Represented by It may correspond to. According to the DASH standard, video data can be divided into video segments (i.e., segments), May represent the length of time in seconds for the video segment. Represents the number of segments, May represent the version number. for example, May represent a chunk video of version j of segment i. Is It can represent the storage capacity of. The data transcoded at the highest bit rate version, ie the original data, can be stored on the hot disk for transcoding other versions. That is, the original data can be stored on a hot disk for transcoding into several versions that are relatively lower than the highest bit rate.

If the client terminal requests a version of the video data corresponding to a specific bit rate, but the requested version of the data is not stored in the hot disk, the transmission control unit 250 is transcoded to a version of the bit rate relatively lower than the requested data Streaming can be transmitted. As such, if the requested data is stored on the cold disk and not stored on the hot disk, the decision unit 220 may determine the lowest bit rate of each video segment in order to service the relatively low bit rate data stored on the hot disk. The corresponding version of the data can be determined as the hot data to be stored in the hot disk. Then, the storage controller 230 may store the video segment corresponding to the hot data on the hot disk. Represents the size of a single disk, May represent the total storage capacity. then, It can be calculated based on. here, May represent the number of hot disks.

Version It can be assumed that the video quality index of. The value can be obtained using the Structural Similarity Index Method (SSIM) tool. Quality of Experiment (QoE) Gain Is Version as It may be predefined to indicate the total QoE achieved for the request for. Then, to the hot disk As shown in Table 1 to indicate whether Can be set to a binary variable. At this time, If is May be determined to be stored on the hot disk. In other words, May be determined as hot data. Otherwise (that is, Is not)), Is determined to be stored on the cold disk, and may be stored on the cold disk. At this time, in the version set of all video segments, Find version An optimization problem for determining whether the data can be stored may be expressed as Equation 1 below.

[Equation 1]

In Equation 1, the optimization problem may correspond to an NP-hard problem as a 0/1 knapsack problem. In the optimization problem, algorithm complexity can be very important when the number of video segments is large. Accordingly, as shown in Table 1, hot data may be determined in consideration of the complexity of the algorithm.

According to Table 1 Each version About It can be defined as. here, May represent the size (ie, capacity) of the disk where the video version j of segment i is to be stored. Then, the storage control unit 230 of the video segment of each version Larger versions can be stored on a hot disk. In this case, the total access probability of the video segment i on the hot disk and the sum of the storage spaces are It can be expressed as. And, May correspond to a binary variable indicating whether video segment i is stored on disk k . The sum of access probabilities for disk k Can be calculated based on .

At this time, since the hot disk is composed of a plurality of disks, the storage control unit 230 may store the hot data on the hot disk to distribute the workload (work lord) evenly according to the popularity. Represents the access probability of video segment i, so that storage control unit 230 stores The sum of can be adjusted evenly. In other words, In order to make the sum equal, a hot disk in which hot data is to be stored among several hot disks can be determined. Of video segment i The problem of optimizing the hot data allocation may be expressed as Equation 2 below.

[Equation 2]

Table 2 below shows a hot data allocation algorithm.

According to Table 2, the storage control unit 230 is a video segment array You can sort in descending order by value. here, Can represent the total access probability for video segment i on the hot disk. The storage controller 230 may sequentially allocate the video segments to the hot disk having the smallest workload of the plurality of hot disks. For example, the storage controller 230 may sequentially allocate the video segments to the plurality of hot disks based on the worst-fit allocation scheme. The total number of requests for hot data stored on the hot disk If you define as Is the video segment for request m, May indicate the disk number for request m. And, May indicate the actual requested version index of request m. Request m To indicate that the first version of the video segment has been sent to the client Using, Seek time, Data rate, version Bitrate , The streaming service time providing the jth version of request m In this case, the streaming service time is It can be calculated based on. That is, the streaming service time providing the j th version of the request m Seek time, data rate, version Bit rate and the length of time for the video segment (ie, segment length, e.g., seconds). It can be calculated based on.

In this case, when the data corresponding to the version requested by the client terminal is not stored in the hot disk, the transmission controller 240 may transmit the video data of a version lower than the requested version to the client terminal. To ensure real time streaming, the total service time spent reading all video segments can be kept below the L seconds, which is the segment length. That is, it should not exceed L seconds. Then, the optimization problem of maximizing the total QoE while ensuring real time streaming in each hot disk can be expressed as Equation 3 below.

[Equation 3]

The optimization problem of Equation 3 is a variation of the multi-selection knapsack problem and may correspond to an NP-hard problem as in Equation 1. Then, a hot disk bandwidth allocation algorithm (ie, bandwidth allocation heuristic algorithm) may be expressed as shown in Table 3 below based on the optimization problem and NP-hard problem of Equation 3.

According to Table 3, the hot disk bandwidth allocation algorithm may allocate the bandwidth of the hot disk by using the parameter. For example, the rate parameter for the jth version of request m May be calculated based on Equation 4 below.

[Equation 4]

In Equation 4, Version Video quality index, Is the actual requested version index of request m, May indicate a streaming service time providing a j-th version of the request m.

According to Equation 4 and Table 3, the index variable for each request silver Can be initialized to I.e. of request m To indicate that the first version of the video segment has been sent to the client May be initialized to the version closest to the requested version among the serviceable versions. For example, if the data corresponding to the version requested from the client terminal is not stored in the hot disk, the Is the version closest to the version requested from the client terminal among the serviceable versions of data stored in the hot disk. Can be initialized to At this time, The value of may be reduced to reduce the bit rate, but the overall service time may not exceed the segment length L seconds. That is, L seconds or less can be maintained.

In Equation 4 above, Corresponds to the molecule of Indicates a decrease in QoE, the denominator May represent an increase in bandwidth. In order to allocate effective bandwidth based on Equation 4, the lowest Select an argument with a value, You can control the value to go low. Like this, as in Table 4 of To the lowest level Select an argument with a value, A sequence of controls to lower the value to a low value is based on the bandwidth requirements of the predefined hot disk. ) May be repeated until it is satisfied.

FIG. 4 is a graph showing QoE and power consumption relationship of 2TB and 3TB in an HVP workload according to an embodiment of the present invention, and FIG. 5 is a LVP workload in an embodiment of the present invention. Is a graph showing the power consumption relationship with QoE of 2TB and 3TB respectively.

4 and 5, a power analysis model may be used to predict power consumption. Is the power required (i.e. power consumption) during the discovery process (i.e. seek mode among multiple assistance modes), Is the power required during the active process (ie, active mode), Is the power required during the idle process (i.e., the idle mode). In this case, the energy consumption required to use L seconds for each process (ie, each mode) can be calculated.

Total seek time is Since the energy required in the search process is It can be calculated based on. And the time spent reading data , The energy consumed to read the data is It can be calculated based on. If no disk activity occurs, the disk spins without reading or seeking. May require power. Accordingly, idle energy silver It can be calculated based on.

Again, in Figures 4 and 5, to determine the relationship between overall QoE and power consumption, five sample videos (e.g., Spiderman, mummies, transformers, planetary escape, Profile a video quality index per second for each resolution, and then SSIM value for, You can extract the value. QoE is usually expressed as an average opinion score (MOS) between 1 (bad) and 5 (good), so the MOS value is determined according to the guidelines shown in Table 4 below. Can be used for

Table 4 may show a mapping table between SSIM and QoE. Based on the guidelines in Table 4 Each of the five sample videos having a resolution of 6 may be transcoded into six different versions. For example, it can be transcoded into six versions: 1600 × 900, 1280 × 720, 1024 × 600, 854 × 480, 640 × 360, and 426 × 240. If you calculate the bitrate using the YouTube HD quality version video bitrate calculator and set the length of each video segment to 2 seconds, the arrival of the client request goes to the Poisson process at the 1 / s arrival rate. Can be modeled. The access probability of each segment follows a Zipf distribution, and [theta] can be set to 0.271 as demonstrated in the actual Video on Demand (VoD) application. Assuming that the video storage of the video storage power management system, which is a server, stores 2000 video files that last from 2 to 3 hours, the client terminal (that is, the user's terminal) plays the video before the video is completed Can be terminated, the actual playback time can be modeled as a Zipf distribution with θ = 0.2. And, the disk parameters of Western Digital, Inc., shown in Table 5 below, may be used.

Cold disks remain in standby mode, which can consume standby power. In addition, the power consumption may be profiled for 24 hours, and the relationship between the QoE and the power consumption may be determined as shown in FIGS. 4 and 5 by considering two workloads as follows.

The first workload is high bit-rate versions are popular, The second workload is a low bit-rate versions are popular (LVP) workload. It may correspond to.

Then, it can be seen based on FIG. 4 and FIG. 5 how the power consumption and the QoE ratio differ according to the number of hot disks compared with storing all the transcoded versions. 4 and 5, instead of storing all transcoded versions of video data on a hot disk, only specific hot data determined according to the number of hot disks are selectively stored on the hot disk and then requested by the client terminal. When streaming the hot data stored in the hot disk by adjusting the version, it can be seen that it effectively reduces the disk power consumption while minimizing the QoE degradation. For example, referring to 410 of FIG. 4, when a 2 TB (Tera Byte) disk is used in the HVP workload, the disk power consumption is reduced by 50% when the QoE 4.7% decreases and 31% when the QoE 0.8% decreases. In addition, referring to 410 and 420, as the number of hot disks increases, more transcoded versions can be stored on the hot disks, thereby increasing the overall QoE, but the power consumption is not significantly affected.

Referring to 510 and 520 of FIG. 5, unlike the HVP workload, in the LVP workload, the QoE degradation is negligible. For example, if the QoE decreases by 1.2% in LVP workload, referring to 520 of FIG. 5, power consumption is reduced by 30% for 3TB disks and 43% for 2TB disks. can confirm.

As such, when compared to 3 TB (Tera Byte) disks, it can be seen that the power reduction effect is greater when using 2 TB disks. This is because when using 2TB disks, more disks are used to store all the files, so selectively storing only hot data on a hot disk and spinning down cold disks is more effective in reducing power consumption than when using many disks. You can check it. According to the graphs of FIGS. 4 and 5, if the simulation results allow 1.5% QoE degradation, selectively storing only hot data on the hot disk, and spinning down the cold disk will store all transcoded versions on the hot disk. It can be seen that the power consumption is reduced by 29% to 46% in comparison with the time.

The method according to the embodiment may be embodied in the form of program instructions that can be executed by various computer means and recorded in a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the media may be those specially designed and constructed for the purposes of the embodiments, or they may be of the kind well-known and available to those having skill in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical media such as CD-ROMs, DVDs, and magnetic disks, 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 not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like.

Although the embodiments have been described by the limited embodiments and the drawings as described above, various modifications and variations are possible to those skilled in the art from the above description. For example, the described techniques may be performed in a different order than the described method, and / or components of the described systems, structures, devices, circuits, etc. may be combined or combined in a different form than the described method, or other components. Or even if replaced or substituted by equivalents, an appropriate result can be achieved.

Therefore, other implementations, other embodiments, and equivalents to the claims are within the scope of the claims that follow.

Claims (10)

In the DASH (Dynamic Adaptive Streaming over HTTP) based video storage power management method,
Transcoding the multimedia data into a plurality of versions corresponding to different bit rates;
Determining hot data to be stored in the hot zone in the video storage divided into a hot zone and a cold zone, for the data transcoded into the plurality of versions;
Storing the determined hot data in a hot zone of the video storage;
Storing data other than the hot data among the data transcoded into the plurality of versions in a cold zone of the video storage; And
Down the cold zone
Including,
Down the cold zone,
Bringing down the cold zone such that hot data stored in the hot zone is selectively accessed
Video storage power management method comprising a. delete In the DASH (Dynamic Adaptive Streaming over HTTP) based video storage power management method,
Transcoding the multimedia data into a plurality of versions corresponding to different bit rates;
Determining hot data to be stored in the hot zone in the video storage divided into a hot zone and a cold zone, for the data transcoded into the plurality of versions;
Storing the determined hot data in a hot zone of the video storage;
Storing data other than the hot data among the data transcoded into the plurality of versions in a cold zone of the video storage; And
Down the cold zone
Including,
When data corresponding to a specific version stored in the cold zone is requested, streaming data corresponding to a transcoded version at a lower bit rate than data stored in the requested cold zone
Video storage power management method further comprising. The method of claim 3,
Data corresponding to a transcoded version at a lower bit rate than data stored in the cold zone indicates data stored in the hot zone
Video storage power management method characterized in that. In the DASH (Dynamic Adaptive Streaming over HTTP) based video storage power management method,
Transcoding the multimedia data into a plurality of versions corresponding to different bit rates;
Determining hot data to be stored in the hot zone in the video storage divided into a hot zone and a cold zone, for the data transcoded into the plurality of versions;
Storing the determined hot data in a hot zone of the video storage;
Storing data other than the hot data among the data transcoded into the plurality of versions in a cold zone of the video storage; And
Down the cold zone
Including,
Determining the hot data,
Determining the original data of the multimedia data corresponding to the highest bit rate among the plurality of versions of the transcoded data, and the data that is transcoded at the lowest bit rate as the hot data.
Video storage power management method characterized in that. In the DASH (Dynamic Adaptive Streaming over HTTP) based video storage power management method,
Transcoding the multimedia data into a plurality of versions corresponding to different bit rates;
Determining hot data to be stored in the hot zone in the video storage divided into a hot zone and a cold zone, for the data transcoded into the plurality of versions;
Storing the determined hot data in a hot zone of the video storage;
Storing data other than the hot data among the data transcoded into the plurality of versions in a cold zone of the video storage; And
Down the cold zone
Including,
Storing in the hot zone,
Uniformly allocating video segments corresponding to the hot data to a plurality of disks constituting the hot zone;
Video storage power management method comprising a. In the DASH (Dynamic Adaptive Streaming over HTTP) based video storage power management system,
A transcoding unit transcoding the multimedia data into a plurality of versions corresponding to different bit rates;
A determination unit for determining hot data to be stored in the hot zone in the video storage divided into a hot zone and a cold zone for the data transcoded into the plurality of versions;
A storage control unit storing the determined hot data in a hot zone of the video storage and storing remaining data except the hot data among the data transcoded into the plurality of versions in a cold zone of the video storage;
A power control unit to down the cold zone; And
As data corresponding to a specific version stored in the cold zone is requested, a transmission control unit for streaming and transmitting data corresponding to a transcoded version at a lower bit rate than data stored in the requested cold zone
Video storage power management system comprising a. delete In the DASH (Dynamic Adaptive Streaming over HTTP) based video storage power management system,
A transcoding unit transcoding the multimedia data into a plurality of versions corresponding to different bit rates;
A determination unit for determining hot data to be stored in the hot zone in the video storage divided into a hot zone and a cold zone for the data transcoded into the plurality of versions;
A storage control unit storing the determined hot data in a hot zone of the video storage and storing remaining data except the hot data among the data transcoded into the plurality of versions in a cold zone of the video storage; And
A power control unit to down the cold zone
Including,
The determination unit,
Determining the original data of the multimedia data corresponding to the highest bit rate among the plurality of versions of the transcoded data, and the data that is transcoded at the lowest bit rate as the hot data.
Video storage power management system, characterized in that. In the DASH (Dynamic Adaptive Streaming over HTTP) based video storage power management system,
A transcoding unit transcoding the multimedia data into a plurality of versions corresponding to different bit rates;
A determination unit for determining hot data to be stored in the hot zone in the video storage divided into a hot zone and a cold zone for the data transcoded into the plurality of versions;
A storage control unit storing the determined hot data in a hot zone of the video storage and storing remaining data except the hot data among the data transcoded into the plurality of versions in a cold zone of the video storage; And
A power control unit to down the cold zone
Including,
The storage control unit,
Uniformly allocating video segments corresponding to the hot data to a plurality of disks constituting the hot zone;
Video storage power management system, characterized in that.