JP7088550B2 - Storage equipment, host computer, management system, and data control method - Google Patents

Description

The present disclosure relates to storage devices, host computers, management systems, and data control methods.

In recent years, various types of data have come to be handled in computer devices, and the amount of data stored in storage is rapidly increasing. Attention is being paid to the data compression function performed when storing such a large amount of data in storage. Data compression is performed by creating a compression attribute volume dedicated to compression in the storage and storing the data there. However, the performance of IO processing is inferior in the access to the compressed attribute volume because the compression / decompression processing works compared to the access to the uncompressed attribute volume. Therefore, in order to prevent performance degradation when accessing a compressed attribute volume, it is effective to store frequently accessed data in the uncompressed attribute volume and store infrequently accessed data in the compressed attribute volume. .. However, it is troublesome for the user to consider whether to arrange the data in the compressed attribute volume or the uncompressed attribute volume, and it is difficult to determine the decision.

Patent Document 1 describes whether or not the device controller compresses the data and saves the data in the non-volatile memory device according to the update frequency level to the data at the timing when the non-volatile memory device receives the data from the host device. It is stated that the judgment is made.

International Publication No. 2016/135954

However, when the process described in Patent Document 1 is executed, the following problem occurs when the update frequency level for data fluctuates greatly in a short period of time. When the update frequency level for data fluctuates greatly in a short period of time, the compressed state and uncompressed state of the data are frequently changed, and there is a problem that the data cannot be stored in the non-volatile memory in the optimum state. be.

An object of the present disclosure is to provide a storage device, a host computer, a management system, and a data control method capable of realizing the optimum arrangement of data in a compressed attribute volume and an uncompressed volume.

The storage device according to the first aspect of the present disclosure uses the score of the data calculated by using the access status to the data in a predetermined period and the analysis result of a plurality of the scores calculated at different timings. When the determination unit for determining whether to store the data in the compressed volume or the uncompressed volume and the current storage location of the data are different from the storage location determined as the storage destination in the determination unit, the data. Is provided with a movement control unit for moving the data to a storage location determined as a storage destination in the determination unit.

The host computer according to the second aspect of the present disclosure is based on a communication unit that collects a score of the data calculated by using the access status of the data in a predetermined period, and a plurality of the scores calculated at different timings. The communication unit includes an analysis unit that analyzes whether the data is stored in the compressed volume or the uncompressed volume, and the communication unit outputs the results analyzed by the analysis unit between the compressed volume and the uncompressed volume. It is transmitted to the storage device that controls the movement of the data in.

The management system according to the third aspect of the present disclosure receives a storage device that transmits a score of the data calculated by using the access status to the data in a predetermined period, and a plurality of the scores calculated at different timings. , A host computer that analyzes whether the data is stored in a compressed volume or an uncompressed volume, and the storage device compresses the data using the score of the data and the analysis result of the score. Determine whether to store in a volume or an uncompressed volume.

In the data control method according to the third aspect of the present disclosure, a score of the data calculated by using the access status of the data in a predetermined period and an analysis result of a plurality of the scores calculated at different timings are obtained. It is used to determine whether to store the data in a compressed volume or an uncompressed volume, and if the current storage location of the data is different from the storage location determined as the storage destination, the data is stored in the determination unit. Move to the storage location determined as the destination.

INDUSTRIAL APPLICABILITY According to the present disclosure, it is possible to provide a storage device, a host computer, a management system, and a data control method capable of realizing the optimum arrangement of data in a compressed attribute volume and an uncompressed volume.

It is a block diagram of the storage apparatus which concerns on Embodiment 1. FIG. It is a block diagram of the storage system which concerns on Embodiment 2. FIG. It is a block diagram of the storage 40 which concerns on Embodiment 2. FIG. It is a figure which shows the table for score calculation which concerns on Embodiment 2. FIG. It is a figure which shows the score table which concerns on Embodiment 2. It is a figure which shows the address translation table which concerns on Embodiment 2. It is a figure which shows the score analysis result table 50 which concerns on Embodiment 2. FIG. It is a block diagram of the storage controller which concerns on Embodiment 2. FIG. It is a figure which shows the flow of the update process of the score calculation table which concerns on Embodiment 2. FIG. It is a figure which shows the flow of the update process of the score table which concerns on Embodiment 2. FIG. It is a figure which shows the flow of the threshold setting process which concerns on Embodiment 2. FIG. It is a block diagram of the host computer which concerns on Embodiment 2. FIG. It is a figure which shows the external score table which concerns on Embodiment 2. FIG. It is an image diagram comparing the predicted score curve and the threshold value with respect to the time required for the second embodiment. It is a figure which shows the flow of the calculation process of the score correction value which concerns on Embodiment 2. FIG. It is a figure which shows the flow of the score analysis processing which concerns on Embodiment 2. It is a figure which shows the flow of the process which concerns on the movement control of the data which concerns on Embodiment 2. It is a block diagram of the disk array apparatus and the host computer which concerns on Embodiment 2. FIG.

(Embodiment 1)
Hereinafter, embodiments of the present invention will be described with reference to the drawings. First, a configuration example of the storage device 10 according to the first embodiment will be described with reference to FIG. The storage device 10 may be a computer device operated by the processor executing a program stored in the memory.

The storage device 10 has a determination unit 11 and a movement control unit 12. The components of the storage device 10 such as the determination unit 11 and the movement control unit 12 may be software or a module whose processing is executed by the processor executing a program stored in the memory. Alternatively, the component of the storage device 10 may be hardware such as a circuit or a chip.

The determination unit 11 determines whether the data is stored in the compressed volume or the uncompressed volume. Specifically, when making a determination, the determination unit 11 uses a data score calculated using the access status to the data in a predetermined period and an analysis result of a plurality of scores calculated at different timings. .. It is assumed that the data is stored in a storage area different from the compressed volume, the uncompressed volume, or the compressed volume and the uncompressed volume in the storage device 10, for example. The storage area different from the compressed volume and the uncompressed volume may be, for example, a cache area. The data score may be calculated higher when, for example, the frequency with which the processor or the like accesses the data, the frequency with which the data is used, or the frequency with which the data is updated is high, as compared with the case where these frequencies are low. good. In other words, frequently used data is set to a higher score than less frequently used data.

The plurality of scores calculated at different timings are, for example, the score of the data calculated using the access status of the data in the first period and the access to the data in the second period different from the first period. It may include the score of the data calculated using the situation. Further, the plurality of scores calculated at different timings may include the scores of the data calculated by using the access status of the data in three or more periods. The predetermined period such as the first period and the second period may be a few seconds, a few minutes, a few hours, a few days, or the like. The analysis of a plurality of scores calculated at different timings is to analyze the scores calculated in a period longer than the period in which one score is calculated.

From this, the determination unit 11 stores the data in the compressed volume or uncompresses it in consideration of the short-term score calculated in a predetermined period and the long-term score including a plurality of short-term scores. It is possible to determine whether to store in the volume. For example, the determination unit 11 may determine whether to prioritize a short-term score or a long-term score according to a specific condition. For example, when the determination unit 11 indicates that the higher the score, the more frequently the processor accesses the data, the determination unit 11 may determine that the data having a score higher than a predetermined threshold value is stored in the uncompressed volume. Further, the determination unit 11 may determine that data having a score lower than a predetermined threshold value is stored in the compressed volume.

When the current storage location of the data is different from the storage location determined as the storage destination by the determination unit 11, the movement control unit 12 moves the data to the storage location determined as the storage destination by the determination unit 11. For example, when the movement control unit 12 determines that the current storage location of the data is a compressed volume and the determination unit 11 determines that the data is stored in an uncompressed volume based on the score, the movement control unit 12 moves the data to the uncompressed volume. Let me. The same applies when the data is currently stored in the uncompressed volume and the determination unit 11 determines that the storage destination is the compressed volume. Further, when the current storage location of the data is a storage area different from the compressed volume and the uncompressed volume, the movement control unit 12 moves the data to the storage location determined by the determination unit 11 as the storage destination.

As described above, the storage device 10 can determine the storage destination of the data by using the short-term score and the long-term score. By considering the long-term score, the storage device 10 can frequently change the compressed state and the uncompressed state of the data even when the update frequency level for the data fluctuates greatly in a short period of time. Can be prevented. As a result, the storage device 10 can store the data in an optimum state.

(Embodiment 2)
Subsequently, a configuration example of the storage system will be described with reference to FIG. The storage system of FIG. 2 has a disk array device 20 and a host computer 60. The disk array device 20 corresponds to the storage device 10 in FIG. The host computer 60 writes data to the disk array device 20 and reads data stored in the disk array device 20. Further, the host computer 60 updates the data stored in the disk array device 20.

The disk array device 20 has a storage controller 30 and a storage 40. The function realized in the storage controller 30 may be executed, for example, by the processor executing a program stored in the memory.

Here, a configuration example of the storage 40 will be described with reference to FIG. The storage 40 has a host utilization area 41 and an internal area 43. The host usage area 41 is an area accessed by the host computer 60 when writing data or reading data. The host utilization area 41 may be used as a cache area, for example. The cache area is used, for example, as an area for temporarily storing data. The host usage area 41 has a user logical volume 42. The user logical volume 42 is used as an area in which data is stored in the host usage area 41. The user logical volume 42 may be constructed, for example, when the host computer 60 writes data to the disk array device 20.

The internal area 43 has an uncompressed attribute volume 44, a compressed attribute volume 45, and a management area 46. The internal area 43 may be, for example, a data storage area configured by using a plurality of physical disks or the like. For example, the uncompressed attribute volume 44, the compressed attribute volume 45, and the management area 46 may be different physical disks. Alternatively, the uncompressed attribute volume 44 and the compressed attribute volume 45 may be different physical disks, and the management area 46 may be the same physical disk as either the uncompressed attribute volume 44 or the compressed attribute volume 45. Alternatively, the uncompressed attribute volume 44, the compressed attribute volume 45, and the management area 46 may be one physical disk.

The uncompressed attribute volume 44 is an area for storing data without compression. On the other hand, the compression attribute volume 45 is an area for compressing and storing data. The management area 46 is an area for managing the score calculation table 47, the score table 48, the address translation table 49, and the score analysis result table 50.

Here, the score calculation table 47 will be described with reference to FIG. The score calculation table 47 manages the number of IOs, the number of writes, the number of cache misses, and the data reduction rate for each data. The number of IOs is the number of times the host computer 60 has accessed the data. For example, when the disk array device 20 receives a SCSI (Small Computer System Interface) command for data 1 from the host computer 60, the number of IOs of data 1 may be incremented.

Further, when the disk array device 20 receives a SCSI command indicating that the data 1 is to be written from the host computer 60, the number of writes of the data 1 may be incremented.

Further, if a cache miss occurs when the host computer 60 accesses the data 1, the number of cache misses of the data 1 may be incremented. For example, when the host computer 60 accesses the cache area in order to read the data 1, if the data 1 does not exist in the cache area, a cache miss occurs.

The data reduction rate is calculated using the data size after compression of the data and the data size before compression. For example, the reduction rate of the data 1 may be calculated as 1- (data size after compression / data size before compression).

The values managed in the score calculation table 47 may be cleared at predetermined intervals. For example, the value managed in the score calculation table 47 may be cleared every period for calculating the score, which is defined in the score table 48 described with reference to FIG.

Subsequently, the score table 48 will be described with reference to FIG. The score table 48 manages the score in each period for each data. For example, the period 0 shown in the score table 48 may be a period from 0 seconds to less than 1 second, assuming that the timing for starting the score calculation is 0 seconds. Further, the period 1 may be a period of 1 second to less than 2 seconds, and the period 2 may be a period of 2 seconds to less than 3 seconds. That is, the score associated in each period may be the score in one second. Alternatively, each period may be determined by using the start time and the end time.

The score in each period is calculated using, for example, the information managed in the score calculation table 47. For example, the score may be calculated as score = (IO count + write count + cache miss count + (IO count x data reduction rate)) / 4.

Subsequently, the address translation table 49 will be described with reference to FIG. The address conversion table 49 manages each data in association with a user logical volume (volume or vol) number, a user logical volume address, a current compression attribute, a later period compression attribute, and an internal volume address. The user logical volume number is a number assigned to identify the user logical volume 42. The user logical volume address may be, for example, the address of the host usage area 41 assigned to the user logical volume 42 in which data is stored. The current compression attribute indicates whether the data is stored in the compressed state or in the uncompressed state. The compression attribute of the later period indicates the result of the storage controller 30 determining whether to store the data in the compressed state or the uncompressed state. The internal volume address may be the address of the internal area 43 indicating the current storage destination of the data.

Subsequently, the score analysis result table 50 will be described with reference to FIG. 7. The score analysis result table 50 manages the function form representing the curve of the analysis score for each data calculated by the host computer 60 and the coefficient indicating the function form in association with each other. The functional form representing the curve of the analysis score may be paraphrased as the functional form used for fitting the analysis score. The function form may be, for example, a linear function, a quadratic function, or the like. When the function form is a quadratic function (ax ² + bx + c), three coefficients are used. Therefore, the respective values of the coefficient 1, the coefficient 2, and the coefficient 3 are managed in the score analysis result table 50.

Subsequently, a configuration example of the storage controller 30 will be described with reference to FIG. The storage controller 30 has a score calculation information storage control unit 31, a score calculation control unit 32, a data movement control unit 33, and a threshold value control unit 34. The storage controller 30 corresponds to the determination unit 11 and the movement control unit 12 in FIG.

The score calculation information storage control unit 31 updates the values managed in the score calculation table 47. The score calculation control unit 32 calculates the score for each predetermined period using the values managed in the score calculation table 47. The score calculation control unit 32 updates the score managed in the score table 48 to the calculated score.

The data movement control unit 33 determines the storage destination of the data by using the instruction message transmitted from the host computer 60 or the score managed in the score calculation table 47. The data movement control unit 33 moves the data when the storage destination of the data determined based on the score or the like is different from the current storage destination of the data. When the data is moved, the data movement control unit 33 updates the current compression attribute and the internal volume address in the address translation table 49. Further, the data movement control unit 33 may update the compression attribute in a later period after determining the data storage destination.

The threshold value control unit 34 sets a threshold value used for comparison with the score when the data movement control unit 33 determines the storage destination of the data. The threshold value control unit 34 does not set the threshold value until the user logical volume 42 is constructed and a certain period of time elapses. For example, the data movement control unit 33 stores all the data in the uncompressed attribute volume 44 until the user logical volume 42 is constructed and a certain period of time elapses. The fixed period may include a period until a fixed capacity is stored in the uncompressed attribute volume 44.

The threshold value control unit 34 sets the threshold value after a certain period of time has elapsed. For example, the threshold control unit 34 calculates the average value of the scores of all the data stored in the user logical volume 42 in a predetermined period. When it is stipulated that the data of the lower A% of the average score (A is a positive value) is stored in the compression attribute volume 45, the score at the boundary between the upper score and the lower score of the average score is specified. Set the value of to the threshold. The high score is a score not included in the low A%. The threshold value control unit 34 may periodically update the threshold value, or may update the threshold value at an arbitrary timing according to an operation by an administrator or the like. Alternatively, the threshold control unit 34 may calculate the average value of the scores of all the data stored in the internal region 43 in a predetermined period.

Subsequently, the flow of the update process of the score calculation table 47 will be described with reference to FIG. First, the score calculation information storage control unit 31 receives a SCSI command related to IO control from the host computer 60 (S11). Next, the host computer 60 increments the number of IOs of the data specified in the SCSI command in the score calculation table 47 (S12). Next, the score calculation information storage control unit 31 determines whether or not the SCSI command is a write command (S13). When the score calculation information storage control unit 31 determines that the SCSI command is a write command, it increments the number of times the data specified in the SCSI command in the score calculation table 47 is written (S14). When the score calculation information storage control unit 31 determines that the SCSI command is not a write command, the score calculation information storage control unit 31 executes the processing of step S15 and subsequent steps without executing the processing of step S14.

Next, the score calculation information storage control unit 31 determines whether or not the data specified in the SCSI command exists in the cache area, that is, whether or not a cache error has occurred (S15). When the score calculation information storage control unit 31 determines that a cache miss has occurred, the score calculation information storage control unit 31 increments the number of cache misses of the data specified by the SCSI command in the score calculation table 47 (S16). When it is determined that no cache error has occurred, the score calculation information storage control unit 31 executes the process of step S17 without executing the process of step S16.

Next, the score calculation information storage control unit 31 calculates the data reduction rate (S17). For example, the score calculation information storage control unit 31 compresses data in the cache area. Further, the score calculation information storage control unit 31 calculates the compression rate as the compression rate = the data size after compression / the data size before compression. Further, the score calculation information storage control unit 31 calculates the data reduction rate with the data reduction rate = 1-compression rate. In order to calculate the data reduction rate, the score calculation information storage control unit 31 deletes the compressed data in the cache area without storing it in the compression attribute volume 45. That is, the score calculation information storage control unit 31 deletes the compressed data in the cache area and releases the cache.

The score calculation information storage control unit 31 updates the data reduction rate in the score calculation table 47 to the calculated data reduction rate. Further, the score calculation information storage control unit 31 may calculate the data reduction rate at a time interval longer than the score calculation interval. The score calculation information accumulation control unit 31 may calculate the data reduction rate at a predetermined timing. In this case, the score calculation information storage control unit 31 may clear the IO count, the write count, and the cache miss count when the score for each predetermined period is calculated by the score calculation control unit 32. Further, the score calculation information storage control unit 31 may maintain the current value of the data reduction rate until the next data reduction rate is calculated.

Subsequently, the flow of the update process of the score table 48 will be described with reference to FIG. First, the score calculation control unit 32 acquires the IO count, the write count, the Cass miss count, and the data reduction rate associated with the data for which the score is calculated from the score table 48 (S21). Next, the score calculation control unit 32 calculates the score of the data to be calculated. The score may be calculated as, for example, score = (IO count + write count + cache miss count + (IO count x data reduction rate)) / 4. Alternatively, the score calculation control unit 32 may give weights to the number of IOs, the number of writes, the number of Cass misses, and the data reduction rate. For example, the score calculation control unit 32 may determine a parameter value to be weighted and a weighting value according to the business content or the like executed by the host computer 60.

Next, the score calculation control unit 32 updates the score in the score table 48 to the calculated score (S23). The score calculation control unit 32 calculates the score for each predetermined period and for each data.

Subsequently, the flow of the threshold value setting process will be described with reference to FIG. First, the threshold value control unit 34 determines whether or not to set the initial threshold value after the user logical volume 42 is constructed (S31). When the threshold value control unit 34 determines that the threshold value is set for the first time, the threshold value control unit 34 sets a predetermined value (S32). For example, when the threshold value is set for the first time, the threshold value control unit 34 may set the threshold value to 0 or set other values. When the threshold is set to 0, the scores of all data exceed the threshold. Therefore, all the data will be stored in the uncompressed attribute volume 44. Further, the threshold value control unit 34 may determine that the threshold value is set for the first time when a value is not set for the current threshold value.

When the threshold value control unit 34 determines that the threshold value is not set for the first time, the threshold value control unit 34 calculates the average value of the scores of all the data stored in the user logical volume 42 in a predetermined period (S33). Next, the threshold value control unit 34 sets the threshold value based on the calculated average value of the scores (S34). For example, when the threshold value control unit 34 is defined to store the data of the lower A% (A is a positive value) of the average value of the score in the compression attribute volume 45, the threshold control unit 34 is the higher score of the average value of the score. The value of the score at the boundary with the lower score is set as the threshold value.

Subsequently, a configuration example of the host computer 60 will be described with reference to FIG. The host computer 60 has an external score table 61, a score analysis control unit 62, and a score correction value calculation control unit 63. The score analysis control unit 62 periodically acquires information managed in the score table 48 of the disk array device 20. The information acquired by the score analysis control unit 62 is managed in the external score table 61.

Here, the external score table 61 will be described with reference to FIG. The external score table 61 of FIG. 13 manages the score related to the data 1. As for the data other than the data 1, the scores are managed in the same manner as in the external score table 61 of FIG.

The external score table 61 manages the cycle and the period in association with each other. The period is, for example, the same as the period in the score table 48. For example, the period may be 1 second or the like. The cycle indicates, for example, a time interval longer than the period, and may be a one-day interval or the like. For example, cycle 0 may indicate the first day, cycle 1 may indicate the second day, and cycle 2 may indicate the third day. The period 0 of the cycle 0 may be, for example, a period from 0 seconds to 1 second on the first day. Further, the period 0 of the cycle 1 may be a period from 0 seconds to 1 second on the second day. That is, the scores for the same period in different cycles may be the scores for the same time zone on different days.

The score analysis control unit 62 uses the score accumulated in the external score table 61 as learning data, analyzes the fluctuation of the score, and predicts the fluctuation of the score. For analysis and prediction of score fluctuations, for example, regression analysis by ridge regularization may be performed, or other methods may be used. The analysis and prediction of score fluctuations using regression analysis by ridge regularization will be described below.

In the case of polynomial fitting using ridge regression, a set of w that minimizes the following equation 1 may be obtained. n is a data number. If t is the elapsed time from the score acquisition start time, S (t _n ) is the score at a certain time, _wi is the coefficient for fitting on the _i -th order curve, and λ is the weighting coefficient, Equation 1 is as follows. It is shown as.

・・・（式１）

... (Equation 1)

This formula is fitted by the steepest descent method or the like. Let the fitted curve be y (t _n, w). The LOOCV method is known as a method for determining λ. The host computer 60 transmits the score information analyzed by using the regression analysis by the polynomial to the disk array device 20 in the format of the score analysis result table 50. The function form indicates the order of which the fitting was performed, and the coefficients after the first coefficient are the fitted coefficients (here, _wi ).

The curve shown by the solid line in FIG. 14 indicates y (t _n, w). FIG. 14 is an image diagram comparing the predicted score curve with respect to time and the threshold value. In FIG. 14, the vertical axis shows the score, and the horizontal axis shows the time.

Next, an example of a method for determining the prediction range will be described. The following procedure predicts whether or not the value taken by the score exceeds (or falls below) the threshold value with a probability of 95% or more.

First, after performing regression analysis, the variance (variability from the regression curve) σ ² is calculated by the following equation 2.

・・・（式２）

... (Equation 2)

It is assumed that the score follows a normal distribution having a variance σ ² derived by the above equation, with the predicted score y (t _n, w) at a certain time as the average. Let z be the value obtained by subtracting the threshold value T from the predicted score y (t _n, w) at a certain time and dividing by the standard deviation σ (the following equation 3).

・・・（式３）

... (Equation 3)

According to the normal distribution table, if this z <1.64, it can be seen that it is in the lower 95% of the normal distribution. Also, if z> -1.64, it is in the top 95%.

・・・（式４）

... (Equation 4)

The right side of the above equation 4 represents the curve shown by the broken line at the lower side of FIG. If the right side exceeds the threshold value T, the score is predicted to exceed the threshold value with a probability of 95% or more. Therefore, regardless of the score calculated by the disk array device 20, the score analysis control unit 62 stores the data in the uncompressed attribute volume 44. The period in which the right side of the equation 4 exceeds the threshold value T is, for example, the periods A1 and A5 in FIG. The periods A1 and A5 may be paraphrased as an absolute uncompressed period.

・・・（式５）

... (Equation 5)

The left side of the above equation 5 represents the curve shown by the broken line on the upper side of FIG. If the right side falls below the threshold T, the score is predicted to fall below the threshold with a probability of 95% or more. Therefore, regardless of the score calculated by the disk array device 20, the score analysis control unit 62 stores the data in the compression attribute volume 45. The period in which the right side of the equation 5 is below the threshold value T is, for example, the period A3 in FIG. The period A3 may be paraphrased as an absolute compression period.

・・・（式６）

... (Equation 6)

The above formula 6 shows the periods A2 and A4 in FIG. The periods A2 and A4 represent the period for adopting the score calculated in the disk array device 20. In this case, the data movement control unit 33 of the disk array device 20 compares the real-time score calculated by the score calculation control unit 32 with the threshold value T.

When the threshold value in FIG. 14 is set to T = 100 and the values are actually applied, the predicted score y at 6 o'clock (6 o'clock) is 75, the predicted score y at 12 o'clock (12 o'clock) is 100, and the predicted score at 18:00. y (18:00) is 170. Assuming that the superiority level is 5% and the standard deviation is σ = 10, 75 + 10 * 1.64 = 91.4 <T at 6 o'clock, so 95% of the score is predicted to be below the threshold. Therefore, the data movement control unit 33 determines that 6 o'clock is a period for surely compressing the data. Similarly, at 12 o'clock, 100-10 * 16.4 <T <100 + 10 * 1.64, so the data movement control unit 33 adopts the storage score. Since T <170-10 * 1.64 at 18:00, the data movement control unit 33 determines that it is a period during which the data is always uncompressed.

As a supplement, the normal distribution centered on the regression curve is described below.

・・・（式７）

... (Equation 7)

Although ridge regression is taken as an example here, other analytical methods may be used for score analysis and prediction. For example, a support vector machine that classifies into three classes of absolutely uncompressed, absolutely compressed, and adopts a storage score may be used, or deep learning may be used.

When the difference between the score and the threshold value is smaller than a predetermined value, the data moves back and forth between the compressed attribute volume 45 and the uncompressed attribute volume 44 due to a slight fluctuation in the score. Therefore, a procedure for solving the problem that data frequently moves between volumes by calculating a score correction value from the business status of the host computer 60 and raising or lowering the score calculated by the storage controller 30 is shown below.

Let F _tot be the number of files used by the host computer 60, F _c be the number of compression-promoting files, and F _nc be the number of non-compression-promoting files. These satisfy the relation of F _c + F _nc = F _tot . The score correction value calculation control unit 63 shown in FIG. 12 examines and counts whether compression is promoted or decompression is promoted based on the usage status of the file. The number of files may be paraphrased as the number of data.

Here, with reference to FIG. 15, the flow of the score correction value calculation process in the score correction value calculation control unit 63 will be described. First, the score correction value calculation control unit 63 determines, for example, whether or not the host computer 60 has write authority for the i-th file (S41).

When the score correction value calculation control unit 63 determines that the host computer 60 has write authority for the i-th file, it determines whether or not the i-th file has been updated within a predetermined period (S42). The score correction value calculation control unit 63 increments F _nc when it is determined that the i-th file has been updated within a predetermined period (S43). Further, when the score correction value calculation control unit 63 determines in step S41 that the host computer 60 does not have write authority for the i-th file, F _c is incremented (S44). Further, the score correction value calculation control unit 63 increments F _c when it is determined in step S42 that the i-th file has not been updated within a predetermined period (S44).

Further, although FIG. 15 shows a process of determining whether to increment F _c or F _nc based on the write permission and the update of the file, other criteria may be further used. .. Alternatively, instead of writing permission and updating the file, it may be determined whether to increment F _c or F _nc based on other criteria.

Next, the score correction value calculation control unit 63 determines whether or not the confirmation of whether to increment F _nc or increment F _c has been completed for all the files in use (S45). When the score correction value calculation control unit 63 determines that the confirmation has not been completed for all the files in use, the process after step S41 is repeated.

When the score correction value calculation control unit 63 determines that the confirmation of all the files in use has been completed, the upper correction coefficient R ₊ for promoting decompression and the lower correction coefficient R for promoting compression are used by using Equation 8. -Calculate ₍ S46).

・・・（式８）

... (Equation 8)

For example, when there are 100 F _tots and (i) F _c = F _nc = 50, the upward correction coefficient R ₊ = 1.5 and the downward correction coefficient R ₋ = 0.5. (ii) When F _c = 1 and F _nc = 99, the upper correction coefficient R ₊ = 1.01 _and the lower correction coefficient R- = 0.99 (1 ≤ R ₊ ≤ 1.5, 0.5 ≤). R _- ≤1).

After calculating the upper correction coefficient R _{+ and} the lower correction coefficient R-, the host computer 60 informs the disk array device 20 whether each file corresponds to decompression promotion or compression promotion, and the upper correction coefficient R. ₊ And the downward correction coefficient R ₋ are transmitted to the disk array device 20 (S47).

The score correction value calculation control unit 63 multiplies the predicted score predicted by the score analysis control unit 62 by the upward correction coefficient R ₊ and the downward correction coefficient R _- to convert the usage data status or state of the host computer 60 into a score. Can be reflected.

Further, the disk array device 20 that has received the upper correction coefficient R _{+ and the lower correction coefficient R- multiplies} the score calculated by the score calculation control unit 32 by the upper correction coefficient R ₊ and the lower correction coefficient R-. The disk array device 20 can reflect the usage data status or state of the host computer 60 by comparing the value obtained by multiplying the upper correction coefficient R _{+ and} the lower correction coefficient R- with the threshold value.

Subsequently, with reference to FIG. 16, the flow of the score analysis process executed by the score analysis control unit 62 will be described. First, the score analysis control unit 62 collects the score table 48 from the disk array device 20 (S51). In other words, the score analysis control unit 62 collects the scores managed in the score table 48 from the disk array device 20.

Next, the score analysis control unit 62 accumulates the collected scores in the external score table 61 (S52). Next, the score analysis control unit 62 uses the score accumulated in the external score table 61 as learning data, analyzes the fluctuation of the score, and predicts the fluctuation of the score (S53). For analysis and prediction of score fluctuations, for example, regression analysis by ridge regularization may be performed, or other methods may be used.

Next, the score analysis control unit 62 transmits the score analysis result to the disk array device 20 (S54). Further, the score analysis control unit 62 may transmit the analyzed score information to the disk array device 20 in the format of the score analysis result table 50.

Subsequently, with reference to FIG. 17, the flow of processing related to data movement control in the data movement control unit 33 will be described. First, the data movement control unit 33 confirms in the score calculation control unit 32 that the calculation of the score of the i-th data has been completed (S61). Next, the data movement control unit 33 determines from the host computer 60 whether or not the i-th data storage destination volume is specified (S62). For example, the data movement control unit 33 confirms whether the absolute compression period and the absolute uncompression period are defined in the analysis result transmitted from the host computer 60 in step S54 of FIG. Further, the data movement control unit 33 confirms whether or not the current timing corresponds to either the absolute compression period or the absolute non-compression period.

When the current timing does not correspond to either the absolute compression period or the absolute non-compression period, the data movement control unit 33 refers to the score of the i-th data in the score table 48 (S63). Further, when the current timing corresponds to either the absolute compression period or the absolute non-compression period, the data movement control unit 33 executes the processing after step S67.

Next, the data movement control unit 33 determines whether or not the correction coefficient for the i-th data has been received (S64). For example, the data movement control unit 33 has information on whether each file corresponds to decompression promotion or compression promotion, as well as an upward correction coefficient R ₊ and a downward correction, transmitted by the host computer 60 in step S47 of FIG. It is determined whether or not the coefficient R ₋ and is received.

When the data movement control unit 33 receives information on whether each file corresponds to decompression promotion or compression promotion, and the upward correction coefficient R ₊ and the downward correction coefficient R-, the i _- th data The value obtained by multiplying the score by the correction coefficient is compared with the threshold value (S65). When the data movement control unit 33 does not receive the information on whether each file corresponds to decompression promotion or compression promotion, and the upper correction coefficient R _{+ and} the lower correction coefficient R-, the i-th data The score is compared with the threshold (S66). Here, the information on whether each file corresponds to the decompression promotion or the compression promotion indicates the information on whether the data corresponds to the decompression promotion or the compression promotion.

Next, the data movement control unit 33 determines whether or not there is movement of the i-th data between volumes (S67). The movement between volumes is a movement between the uncompressed attribute volume 44 and the compressed attribute volume 45. Next, when the data movement control unit 33 determines that there is movement of the i-th data between volumes, the data movement control unit 33 moves the i-th data to the storage destination volume (S68). When the data movement control unit 33 determines that there is no movement of the i-th data between volumes, the data movement control unit 33 executes the processes after step S69.

Next, the data movement control unit 33 determines whether or not the movement determination for all the data has been completed (S69). When it is determined that the movement determination for all the data has not been completed, the data movement control unit 33 executes the processing after step S61 for the i + 1th data. Further, when the data movement control unit 33 determines that the movement determination for all the data has been completed, the data movement control unit 33 ends the process.

As described above, the disk array device 20 according to the second embodiment requires the user to determine whether to store the data in the compressed attribute volume or the uncompressed attribute volume based on the compression rate and the frequency of use depending on the time zone. do not do. By executing analysis using the score, the disk array device 20 arranges data so as to optimize the used capacity and the performance, and can reduce the used capacity of the disk while preventing the performance deterioration due to the compression process. can.

Further, the host computer 60 can correct the data arrangement determination of the disk array device 20 from the IO tendency according to the time zone and the situation of the host computer 60. As a result, short-time movement and high-frequency movement between the compressed attribute volume and the uncompressed attribute volume are suppressed, and the performance of the disk array device 20 can be stabilized.

Further, in the above embodiment, an example in which the host computer 60 performs score analysis has been described. However, when the disk array device 20 has the function of the host computer 60, the disk array device 20 analyzes the score. It may be carried out.

FIG. 18 is a block diagram showing a configuration example of the disk array device 20 and the host computer 60. Referring to FIG. 18, the disk array device 20 and the host computer 60 include a network interface 1201, a processor 1202, and a memory 1203. The network interface 1201 is used to communicate with other network node devices constituting the communication system. The network interface 1201 may include, for example, a network interface card (NIC) compliant with the IEEE 802.3 series. Alternatively, network interface 1201 may be used to perform wireless communication. For example, the network interface 1201 may be used for wireless LAN communication or mobile communication specified in 3GPP (3rd Generation Partnership Project).

The processor 1202 reads software (computer program) from the memory 1203 and executes it to perform the processing of the disk array device 20 and the host computer 60 described by using the flowchart in the above-described embodiment. The processor 1202 may be, for example, a microprocessor, an MPU (Micro Processing Unit), or a CPU (Central Processing Unit). Processor 1202 may include a plurality of processors.

The memory 1203 is composed of a combination of a volatile memory and a non-volatile memory. Memory 1203 may include storage located away from processor 1202. In this case, processor 1202 may access memory 1203 via an I / O interface (not shown).

In the example of FIG. 18, memory 1203 is used to store software modules. The processor 1202 can perform the processing of the disk array device 20 and the host computer 60 described in the above-described embodiment by reading these software modules from the memory 1203 and executing them.

As described with reference to FIG. 18, each of the processors included in the disk array device 20 and the host computer 60 contains one or more programs including instructions for causing the computer to perform the algorithm described with reference to the drawings. Run.

In the above example, the program can be stored and supplied to the computer using various types of non-transitory computer readable medium. Non-temporary computer-readable media include various types of tangible storage media. Examples of non-temporary computer-readable media include magnetic recording media, magneto-optical recording media (eg, magneto-optical disks), CD-ROMs (Read Only Memory), CD-Rs, CD-R / Ws, and semiconductor memories. The magnetic recording medium may be, for example, a flexible disk, a magnetic tape, or a hard disk drive. The semiconductor memory may be, for example, a mask ROM, a PROM (Programmable ROM), an EPROM (Erasable PROM), a flash ROM, or a RAM (Random Access Memory). The program may also be supplied to the computer by various types of transient computer readable media. Examples of temporary computer readable media include electrical, optical, and electromagnetic waves. The temporary computer-readable medium can supply the program to the computer via a wired communication path such as an electric wire and an optical fiber, or a wireless communication path.

The present disclosure is not limited to the above embodiment, and can be appropriately modified without departing from the spirit.

Some or all of the above embodiments may also be described, but not limited to:
(Appendix 1)
The score of the data calculated using the access status to the data in a predetermined period and the analysis result of a plurality of the scores calculated at different timings are used to convert the data into either a compressed volume or an uncompressed volume. A judgment unit that determines whether to store, and
When the current storage location of the data is different from the storage location determined as the storage destination by the determination unit, the movement control unit for moving the data to the storage location determined as the storage destination by the determination unit is provided. Storage device.
(Appendix 2)
The score of the data is calculated by using the access frequency to the data, the number of times to write to the data, the cache miss to the data, and the compression rate when the data is compressed, according to Appendix 1. Storage device.
(Appendix 3)
The determination unit
When the fluctuation of the score is analyzed, the data is stored in the uncompressed volume in the first period in which the range predicted as the possible value of the score of the data exceeds a predetermined value, and the data is stored in the uncompressed volume. The storage device according to Appendix 1 or 2, wherein it is determined that the data is stored in the compressed volume in the second period in which the range predicted as the possible value of the score does not exceed a predetermined value.
(Appendix 4)
The determination unit
When the variation in the score is analyzed, the data is compared with the predetermined value in the third period in which the predetermined value is included in the range predicted as the possible value of the score of the data. The storage device according to Appendix 3, which determines whether to store the data in a compressed volume or an uncompressed volume based on the result.
(Appendix 5)
The determination unit
In the first period and the second period, it is determined whether the data is stored in the compressed volume or the uncompressed volume without using the comparison result between the data and the predetermined value. The storage device according to 3 or 4.
(Appendix 6)
The determination unit
When it is determined whether the data should be stored in the compressed volume or the uncompressed volume according to a predetermined standard, the data is stored in the compressed volume or the uncompressed volume. The storage device according to any one of Supplementary note 1 to 5, which corrects the score of data.
(Appendix 7)
The storage according to any one of Supplementary note 1 to 6, further comprising a communication unit that transmits a plurality of the scores calculated at different timings to a host computer and receives analysis results of the plurality of the scores from the host computer. Device.
(Appendix 8)
A communication unit that collects the score of the data calculated using the access status of the data in a predetermined period, and
It is provided with an analysis unit that analyzes whether the data is stored in a compressed volume or an uncompressed volume based on a plurality of scores calculated at different timings.
The communication unit
A host computer that transmits the results analyzed by the analysis unit to a storage device that controls the movement of the data between the compressed volume and the uncompressed volume.
(Appendix 9)
The analysis unit
It is analyzed whether or not the range predicted as a possible value of the score of the data exceeds a predetermined value.
The communication unit
In the first period in which the range predicted as the possible value of the score of the data exceeds the predetermined value, a message instructing the storage of the data in the uncompressed volume is transmitted to the storage device. In the second period in which the range predicted as the possible value of the score of the data does not exceed the predetermined value, a message instructing the storage of the data in the compressed volume is transmitted to the storage device. The host computer according to Appendix 8.
(Appendix 10)
A storage device that transmits the score of the data calculated using the access status of the data in a predetermined period, and
A host computer that receives a plurality of the scores calculated at different timings and analyzes whether the data is stored in a compressed volume or an uncompressed volume.
The storage device is
A management system that determines whether to store the data in a compressed volume or an uncompressed volume using the score of the data and the analysis result of the score.
(Appendix 11)
The storage device is
It is analyzed whether or not the range predicted as the possible value of the score of the data exceeds a predetermined value, and the range predicted as the possible value of the score of the data exceeds the predetermined value. In the first period, a message instructing that the data is stored in the uncompressed volume is transmitted to the storage device, and the range predicted as a possible value of the score of the data exceeds the predetermined value. The management system according to Appendix 10, which sends a message instructing that the data to be stored in the compressed volume to the storage device in the second period.
(Appendix 12)
The score of the data calculated using the access status to the data in a predetermined period and the analysis result of a plurality of the scores calculated at different timings are used to convert the data into either a compressed volume or an uncompressed volume. Determine if you want to store it
A data control method in a storage device that moves the data to a storage location determined as a storage destination when the current storage location of the data is different from the storage location determined as the storage destination.
(Appendix 13)
Collect the score of the data calculated by using the access status of the data in a predetermined period.
Based on the plurality of scores calculated at different timings, it is analyzed whether the data is stored in the compressed volume or the uncompressed volume.
A communication method in a host computer that transmits the analyzed result to a storage device that controls the movement of the data between the compressed volume and the uncompressed volume.
(Appendix 14)
The score of the data calculated using the access status to the data in a predetermined period and the analysis result of a plurality of the scores calculated at different timings are used to convert the data into either a compressed volume or an uncompressed volume. Determine if you want to store it
A program that causes a computer to move the data to a storage location determined as a storage destination by the determination unit when the current storage location of the data is different from the storage location determined as the storage destination.
(Appendix 15)
Collect the score of the data calculated by using the access status of the data in a predetermined period.
Based on the plurality of scores calculated at different timings, it is analyzed whether the data is stored in the compressed volume or the uncompressed volume.
A program that causes a computer to send the analyzed result to a storage device that controls the movement of the data between the compressed volume and the uncompressed volume.

10 Storage device 11 Judgment unit 12 Movement control unit 20 Disk array device 30 Storage controller 31 Score calculation information storage control unit 32 Score calculation control unit 33 Data movement control unit 34 Threshold control unit 40 Storage 41 Host usage area 42 User logical volume 43 Internal Area 44 Uncompressed attribute volume 45 Compressed attribute volume 46 Management area 47 Score calculation table 48 Score table 49 Address conversion table 50 Score analysis result table 60 Host computer 61 External score table 62 Score analysis control unit 63 Score correction value calculation control unit

Claims (10)

The data using the score of the data calculated by using the access status to the data in a predetermined period and the compression rate when the data is compressed, and the analysis result of a plurality of the scores calculated at different timings. A determination unit for determining whether to store the data in a compressed volume or an uncompressed volume, and
When the current storage location of the data is different from the storage location determined as the storage destination by the determination unit, the movement control unit for moving the data to the storage location determined as the storage destination by the determination unit is provided. Storage device. The score of the data is calculated according to claim 1, wherein the score of the data is calculated by using the access frequency to the data, the number of times to write to the data, the cache miss to the data, and the compression rate when the data is compressed. Storage device. The determination unit
When the fluctuation of the score is analyzed, the data is stored in the uncompressed volume in the first period in which the range predicted as the possible value of the score of the data exceeds a predetermined value, and the data is stored in the uncompressed volume. The storage device according to claim 1 or 2, wherein it is determined that the data is stored in the compressed volume in the second period in which the range predicted as the possible value of the score does not exceed a predetermined value. The determination unit
When the variation in the score is analyzed, the data is compared with the predetermined value in the third period in which the predetermined value is included in the range predicted as the possible value of the score of the data. The storage device according to claim 3, wherein it is determined whether the data is stored in a compressed volume or an uncompressed volume based on the result. The determination unit
A claim for determining whether to store the data in a compressed volume or an uncompressed volume without using the comparison result between the data and the predetermined value in the first period and the second period. Item 6. The storage device according to Item 3 or 4. The determination unit
When it is determined whether the data should be stored in the compressed volume or the uncompressed volume according to a predetermined standard, the data is stored in the compressed volume or the uncompressed volume. The storage device according to any one of claims 1 to 5, which corrects the score of the data. The invention according to any one of claims 1 to 6, further comprising a communication unit that transmits a plurality of the scores calculated at different timings to a host computer and receives analysis results of the plurality of the scores from the host computer. Storage device. A communication unit that collects the score of the data calculated by using the access status to the data in a predetermined period and the compression rate when the data is compressed .
It is provided with an analysis unit that analyzes whether the data is stored in a compressed volume or an uncompressed volume based on a plurality of scores calculated at different timings.
The communication unit
A host computer that transmits the results analyzed by the analysis unit to a storage device that controls the movement of the data between the compressed volume and the uncompressed volume. A storage device that transmits the score of the data calculated by using the access status to the data in a predetermined period and the compression rate when the data is compressed , and
A host computer that receives a plurality of the scores calculated at different timings and analyzes whether the data is stored in a compressed volume or an uncompressed volume.
The storage device is
A management system that determines whether to store the data in a compressed volume or an uncompressed volume using the score of the data and the analysis result of the score. The data using the score of the data calculated by using the access status to the data in a predetermined period and the compression rate when the data is compressed, and the analysis result of a plurality of the scores calculated at different timings. Determine whether to store the data in a compressed volume or an uncompressed volume.
A data control method in a storage device that moves the data to a storage location determined as a storage destination when the current storage location of the data is different from the storage location determined as the storage destination.

Images