Data copy system apparatus and method

The present invention relates to systems, devices, methods and programs for managing data, and in particular to systems, devices, and methods for copying data.

Efficient data copy / move solutions are an important aspect of data management and storage. Copying data may involve moving data from one location to another, such as defragmenting or backing up data. The solution for copying data is the traditional input / output (I / O) operation in which the host system reads and writes data from and to the data storage device. Is included. While these techniques are useful for copying data, they create an unwanted data copy overhead due to the required input / output operations.

Other data copy solutions include IBM's FlashCopy (R), where a single request is sent from the host system to the data storage subsystem, and the data storage subsystem performs the full copy operation. Includes fast replication of. Fast replication does not require multiple input / output operations and is often considered faster and more efficient than input / output techniques. Fast replication can reduce the data copy overhead created by input / output operations, but there are certain problems.

For example, if the host system sends a fast replication request for data that is already in a fast replication relationship, the fast replication request cannot be executed. Therefore, the host system uses traditional I / O operation because waiting for data to be available for fast replication can take longer than copying data with traditional I / O operation. Copy the data. In other words, the data is copied through traditional I / O operations because the host system has no way of estimating when the data will be available for fast replication.

<p> In summary, due to the problems mentioned above, the fast replication operation is still underutilized.</p>

<p> From the above considerations, there is a need for enhanced systems, equipment, and methods for copying data. Beneficially, these systems, devices, and methods provide fast replication by providing a solution for estimating when data will be available for fast replication and for sending additional fast replication requests. It will maximize the data copied through.</p><p> The present invention has been developed in response to current technologies, in particular to the problems and needs of the art that are not completely solved by currently available data copy solutions. Therefore, the present invention maximizes the data copied through fast replication by providing a solution for estimating when the data will be available for fast replication and for sending additional fast replication requests. It has been developed to provide equipment, systems, and methods for copying data to the limit.</p><p> In one aspect of the invention, the system for copying data sends a data storage subsystem capable of supporting fast replication and a fast replication request for the data to the data storage subsystem for fast replication. If the request is not accepted, estimate the expected latency for the pending fast replication operation, wait for the expected latency to expire, send additional fast replication requests for the data to the data storage subsystem, and perform additional fast replication. Includes a host system that records data as moved if the request is granted. In one embodiment, if the expected latency exceeds the expected duration to perform a traditional I / O operation, the host system moves data through the traditional I / O operation.</p><p> In some embodiments, if data is moved through traditional I / O operations, the host system records the data as being moved. In certain embodiments, the fast replication request can include a request to move data to a location local to the data storage subsystem. In other embodiments, the fast replication request can include a request to move data to a non-local location with respect to the data storage subsystem. In one embodiment, the host system adjusts the expected latency according to the workload of the data storage subsystem. In one embodiment, the host removes data from the list of data that will be moved and records the time the data was moved.</p><p> In one aspect of the invention, the device for copying data is a communication module that sends a fast replication request for the data to the data storage subsystem, and a pending fast replication operation if the fast replication request is not granted. Includes a synchronization module that estimates the expected latency for and a data management module that waits for the expected latency to expire. The communication module also sends additional fast replication requests for the data to the data storage subsystem, and the data management module records the data as moved if additional fast replication requests are granted. In some embodiments, the I / O module moves data through the traditional I / O operation if the expected latency exceeds the expected time required to perform the traditional I / O operation.</p><p> The method of the invention for copying data is also presented. The methods in the disclosed embodiments substantially include the operations necessary to perform the functions presented above with respect to the described systems and devices. In one embodiment, the method is to send a fast replication request for the data to the data storage subsystem, to estimate the expected wait time for the pending fast replication operation if the fast replication request is not accepted, the expected wait. This includes waiting for the expiration of time, sending additional fast replication requests for the data to the data storage subsystem, and recording the data as moved if additional fast replication requests are granted. In certain embodiments, the method also includes moving data through the traditional I / O operation if the expected wait time exceeds the expected time required to perform the traditional I / O operation.</p><p> From the first aspect, the present invention provides a system for copying data, which system includes at least one data storage subsystem configured to support fast replication and high speed with respect to the data. Sending replication requests to at least one data storage subsystem, estimating the expected latency for pending fast replication operations if fast replication requests are not accepted, waiting for the expected latency to expire, data It is configured to send additional fast replication requests for to at least one data storage subsystem and, if additional fast replication requests are granted, record the data as moved. It is equipped with a host system.</p><p> Preferably, the invention is further configured to move data through conventional I / O operation if the expected latency exceeds the expected duration for conventional I / O operation. Provide a system.</p><p> Preferably, the present invention provides a system in which the host system is further configured to record the data as if it had been moved through conventional I / O operations.</p><p> Preferably, the present invention provides a system in which the host system is further configured to adjust the expected latency according to the workload of at least one data storage subsystem.</p><p> Preferably, the invention is further configured to record the data as moved by the host system by removing the entries representing the data from the list of entries representing the data to be moved. Provide a system.</p><p> Preferably, the invention provides a system in which the host system is further configured to record the time the data was moved.</p><p> Preferably, the present invention provides a system in which a fast replication request is a request for moving data to a non-local location with respect to at least one data storage subsystem.</p><p> In view of the second aspect, the present invention provides a device for copying data, the device comprising a communication module configured to send a fast replication request for data to a data storage subsystem. A synchronization module configured to estimate the expected latency for a pending fast replication operation if a fast replication request is not accepted, a data management module configured to wait for the expected latency to expire, and data. A communication module further configured to send additional fast replication requests to the data storage subsystem, and further configured to record data as moved if additional fast replication requests are granted. It has a data management module.</p><p> Preferably, the invention is further configured to move data via conventional I / O operation if the expected latency exceeds the expected time required for conventional I / O operation. Provide the device.</p><p> Preferably, the invention provides a device in which a data management module is further configured to record data as it has been moved.</p><p> Preferably, the invention provides a device in which the synchronization module is further configured to adjust the expected latency according to the workload of the data storage subsystem.</p><p> Preferably, the invention is further configured such that the data management module records the data as moved by removing the entries representing the data from the list of entries representing the data to be moved. Provide the device.</p><p> Preferably, the invention provides a device in which a data management module is further configured to record the time the data has been moved.</p><p> From a third aspect, the present invention provides a signal-bearing medium that reliably records a program of machine-readable instructions that can be executed by a digital processor to perform an operation for copying data. This operation sends a fast replication request for data to the data storage subsystem, estimates the expected wait time for a pending fast replication operation if the fast replication request is not accepted, and the expiration of the expected latency. Includes waiting for, sending additional fast replication requests for data to the data storage subsystem, and recording the data as moved if additional fast replication requests are granted.</p><p> Preferably, the present invention further comprises moving data through conventional I / O operation when the expected latency exceeds the expected time required for conventional I / O operation. provide.</p><p> Preferably, the invention provides a signal-carrying medium, further comprising recording the data as if the data were moved, if the operation was moved.</p><p> Preferably, the invention provides a signal-carrying medium, further comprising adjusting the expected latency according to the workload of the data storage subsystem.</p><p> Preferably, the present invention provides a signal-carrying medium comprising recording data as moved involves removing an entry representing the data from a list of entries representing the data to be moved.</p><p> Preferably, the invention provides a signal-carrying medium, further comprising recording the time the data has been moved.</p><p> From a fourth aspect, the invention provides a method for copying data, which method sends a fast replication request for the data to the data storage subsystem, which is not allowed. If you estimate the expected wait time for a pending fast replication operation, or if the expected wait time exceeds the expected duration for a traditional I / O operation, move the data through the traditional I / O operation. If the expected latency does not exceed the expected duration for traditional I / O operation, wait for the expected latency to expire, send additional fast replication requests for data to the data storage subsystem, and add Includes recording data as moved if a fast replication request is granted. In certain embodiments, the method comprises moving data through a conventional I / O operation if the expected waiting time exceeds the expected time required to perform the conventional I / O operation.</p><p> From a fifth aspect, the present invention is a digital computer comprising a software code portion for execution for executing the present invention, as described above, when the computer program product is executed on the computer. Provides computer program products that can be loaded into the internal memory of.</p><p> Reference to features, advantages, or similar terms throughout the specification should indicate that all features and benefits achievable using the present invention are in any single embodiment of the present invention. Or it does not suggest that there is. Rather, the phrase referring to features and benefits is understood to mean that the particular features, benefits, or properties described in connection with an embodiment are included in at least one embodiment of the invention. Yeah. Accordingly, discussions of features and advantages, as well as similar terms, may, but are not necessarily limited to, refer to the same embodiments throughout the specification.</p><p> Hereinafter, embodiments of the present invention will be described in detail as mere examples with reference to the accompanying drawings.</p>

<figref num="1">FIG. 6 is a schematic block diagram showing an embodiment of a data copy system according to a preferred embodiment of the present invention.</figref><figref num="2">FIG. 6 is a schematic block diagram showing an embodiment of a host system according to a preferred embodiment of the present invention.</figref><figref num="3">It is a flow chart which shows one Embodiment of the method for copying data according to a preferable embodiment of this invention.</figref><figref num="4">It is a flow chart which shows one Embodiment of the method for copying data according to a preferable embodiment of this invention.</figref>

Many of the functional units described herein are labeled as modules to specifically emphasize their implementation independence. For example, the module can be implemented as a custom VLSI circuit or as a hardware circuit with off-the-shelf semiconductors such as gate arrays, logic chips, transistors, or other separate components. Modules can also be implemented within programmable hardware devices such as field programmable gate arrays, programmable array logic, and programmable logic devices.

Modules can also be implemented in software for execution by various types of processors. The executable code identification module can include, for example, one or more physical or logical blocks of computer instructions that can be organized as objects, procedures, or functions. Nevertheless, the executable code of the identification module does not have to be physically put together, but when logically combined, it has a module and is different, achieving the module's defined purpose. It can be equipped with different types of instructions stored in a location.

In fact, a module of executable code can be a single instruction or multiple instructions, across several different code segments, in different programs, in different processors, and in some memory devices. It can even be dispersed throughout. Similarly, operational data can be identified and exemplified within modules herein, embodied in any suitable form, and organized within any suitable type of data structure. is there. Operable data can be collected as a single data set or distributed across different locations, including across different storage devices, and at least in part as a system or just an electronic signal. It can exist on the network.

References to "one embodiment," "embodiment," or similar terminology throughout this specification have specific features, structures, or properties that are described in connection with that embodiment, but at least one of the present invention. Means included in one embodiment. Thus, throughout the specification, the appearance of phrases "in one embodiment," "in an embodiment," and similar terms may all refer to the same embodiment, but not necessarily.

References to computer-readable media can take any form that can generate a signal, generate a signal, or run a program of machine-readable instructions on a digital processor. .. Computer-readable media include transmission lines, compact discs, digital video discs, magnetic tapes, Bernoulli drives, magnetic discs, holographic discs or tapes, punched cards, flash memory, magnetoresistive memory, integrated circuits, etc. Alternatively, it can be embodied by another digital processing device memory device.

Moreover, the described features, structures, or properties of the present invention can be combined in any suitable manner in one or more embodiments. However, engineers in the relevant field will understand that the present invention is feasible without one or more specific details, or using other methods, components, materials, and the like. In other instances, well-known structures, materials, or operations are not illustrated or described in detail so as not to obscure aspects of the invention.

FIG. 1 is a schematic block diagram showing an embodiment of a data copy system 100 according to the present invention. The indicated system 100 includes a host system 110 and one or more data storage subsystems 120. The components of System 100 work together to improve data copy efficiency.

In one embodiment, the host system 110 sends a fast replication request for the data 122 to the data storage subsystem 120. The data storage subsystem 120 can include any device capable of supporting fast replication and / or FlashCopy (R) relationships and / or operations. If the fast replication request is granted by the data storage subsystem 120, the host system 110 may record the data 122 as being moved. If the fast replication request is not accepted, the host system 110 can estimate the expected wait time for the pending fast replication operation. In some embodiments, the expected latency is intended to predict when the data 122 will be available for fast replication.

In one embodiment, the host system 110 can wait for the expected latency to expire and then immediately send an additional fast replication request to the data storage subsystem 120. If an additional fast replication request is granted, the host system 110 may record the data 122 as being moved. In some embodiments, if no additional fast replication request is granted, the host system 110 estimates the expected wait time and waits for the wait time to expire until the fast replication request for data 122 is granted. The operation of sending another high-speed replication request is circulated again. Interspersing fast replication requests with latency increases efficiency by reducing the probability that unsuccessful fast replication requests will be sent to the data storage subsystem 120.

In one embodiment, the host system 110 adjusts the expected latency according to the current workload of the data storage subsystem 120. In some embodiments, the host system 110 moves the data 122 through the traditional I / O operation if the expected latency exceeds the expected duration for the traditional I / O operation. In this way, the invention can determine and request the most efficient data copy solution, depending on the workload of the data storage subsystem 120.

FIG. 2 is a schematic block diagram showing an embodiment of the host system 200 according to the present invention. The indicated host system 200 includes a communication module 210, a synchronization module 220, a data management module 230, and an input / output module 240. The modules of the host system 200 are collaboratively adapted to facilitate an efficient data copy solution.

In one embodiment, the communication module 210 sends a fast replication request for data to a data storage subsystem (not shown). The fast replication request can include commands for establishing a fast replication relationship or FlashCopy (R) operation. Data related to fast replication or behavior can be organized by atomic movements such as bytes, words, blocks, tracks, sectors, extents, etc. In some embodiments, the fast replication request includes a request to copy data to and from a location local to the data storage subsystem. In other embodiments, the fast replication request includes a request to copy data to or from a location that is not local to the data storage subsystem.

The data management module 230 can record data as if it has been moved if a fast replication request is granted. In one embodiment, recording the data as moved removes the data-representing entry from the list of data-representing entries to be moved, and records the time when the fast replication request was granted. Can include doing. In this way, the host system 200 can maintain a current and organized list of data that will be copied.

If the fast replication request is not accepted, the synchronization module 220 can estimate the expected wait time for the pending fast replication operation. In one embodiment, the communication module 210 sends an additional fast replication request to the data storage subsystem after the expected latency has expired. If an additional fast replication request is granted, the data management module 230 can record the data as if it had been moved. If the additional fast replication request is unsuccessful, the host system 200 can iterate some or all of the above operations to send another fast replication request after the expected latency.

In some embodiments, the input / output module 240 moves data through the traditional I / O operation if the expected latency exceeds the expected time required for the traditional I / O operation. In such an embodiment, the data management module 230 can record the data as moved and the time of exercise. In some embodiments, the data management module 230 records data as being moved by conventional I / O operation, as opposed to fast replication. Therefore, the present invention determines the most efficient method for copying data, issues a fast copy request or conventional input / output command to execute it, and of the copied data. It is possible to record the situation.

In some embodiments, the synchronization module 220 can adjust the expected latency according to the workload of the data storage subsystem. For example, if the data storage subsystem 120 is experiencing a workload reduction, the synchronization module 220 can reduce the expected latency for future fast replication requests. Similarly, if the data storage subsystem 120 is experiencing an increase in workload, the synchronization module 220 can increase the expected latency for future fast replication requests. And the system 100 Therefore, by the executable adjustable wait time for fast replication request, to minimize the number of communications sent to and from the host system 110 and the data storage subsystem 120 To do.

FIG. 3 is a flow chart showing an embodiment of Method 300 for copying data according to the present invention. The indicated method 300 includes an action 310 for sending a fast replication request, an action 320 for determining whether a fast replication request has been granted, an action 330 for estimating the expected wait time for a pending fast replication request, and an expected wait. Action 340 that waits for the expiration of time, action 350 that sends another high-speed replication request, action 360 that determines whether an additional high-speed replication request is accepted, and if a high-speed replication request is granted, it is assumed that it has been moved. Includes operation 370 to record data. The operation of method 300 shows one method for efficiently copying data.

Sending a fast replication request 310 can include host system 110 sending a fast replication request for data 122 to the data storage subsystem 120. Determining whether a fast replication request is accepted 320 may include the data storage subsystem 120 performing the fast replication request. If a fast replication request is granted, the host system 110 can record the data as if it had been moved 370. In certain embodiments, recording the data as being moved 370 may include removing an entry representing the data from the list of entries representing the data to be moved. The rejected fast replication request may include a scenario in which the data corresponding to the fast replication request is already in a fast replication relationship.

If the fast replication request is not accepted 320, the host 110 can estimate the expected wait time for the pending fast replication operation 330. Estimating the expected latency 330 can include multiplying the amount of data that will be moved by the selected exponential value. Waiting for the expected wait time to expire 340 can include that the host system 110 does not send a fast replication request to the data storage subsystem 120. With the expected latency expired 340, host 110 can send other fast replication requests to the data storage subsystem 120 350.

Determining whether a fast replication request is accepted 360 can include the data storage subsystem 120 performing the fast replication request. If fast replication is allowed, the host system 110 can record the data as if it had been moved 370. If the fast replication request is not accepted, the host system can re-estimate the expected latency for the next fast replication request.

FIG. 4 is a flow chart showing an embodiment of Method 400 for copying data according to the present invention. The method 400 shown is via traditional I / O actions: action 405 to determine if there is data to move, action 425 to send a fast replication request, action 430 to send an additional fast replication request. Includes multiple actions, including actions 435 that move data as appropriate. The operation of Method 400 demonstrates one of the many possible methods of the invention for copying or moving data.

Determining if there is data to move 405 includes having host system 110 access a list, array, or table of entries (not shown) that represent the data to be moved. Can be done. To illustrate FIG. 4, the table of entries contains a first entry that represents the first data track on the data storage subsystem. However, in certain embodiments or embodiments, the list of entries includes multiple entries that need to be moved or copied. If the host system 110 determines that the list of entries contains the first entry, the host system 110 can determine whether a fast replication request was attempted for the first entry 410.

If the fast replication request is not attempted, the host system 110 may send a fast replication request for the first data track to the data storage subsystem 120425. If a fast replication request is granted 450, the host system can record the first data track as moved 470. In certain embodiments, this can include removing the first entry from the list of entries and notifying the time of travel.

If the fast replication request is not granted 450, the host system 110 can estimate the expected latency, which estimates when the first data track becomes available for movement 445. In certain embodiments, this can include estimating when the first data track will leave the fast replication relationship. In some embodiments, this is the data unit that will be moved to represent the average time for the data storage subsystem 120, with an average workload for moving one data unit. The number can include multiplying by the selected wait time index.

After estimating the expected latency 445, the host system can record a fast replication request 480 as failing with respect to the first entry. After recording the failure of fast replication 480, or after recording the data as moved 470, the host system 110 can re-determine if there is data to move 405. Assuming that the fast replication attempt was not allowed 450, the host system 110 would discover that the first entry was still in the list of entries. Therefore, the host system 110 will determine if a fast replication attempt has been performed for the first entry 410.

Since the first replication request was attempted for the first entry, host system 110 becomes 415 to determine if the expected wait time has expired. When the expected latency expires, the host system 110 can send other fast replication requests to the data storage subsystem 120 430. If a fast replication request is granted 455, the host system 110 can reduce the time index 475, assuming that the data storage subsystem is experiencing a lower average workload. After reducing the expected latency index 475, the host system 110 can record the data as moved.

Assuming that the fast replication request was not granted 455, the host system assumes that the failure of the fast replication attempt is due to the data storage subsystem 120 experiencing an increased workload. Based on, the expected latency index can be increased 460. The host system 110 can then estimate the expected latency for other fast replication requests 465. In some embodiments, estimating the expected latency 465 can include multiplying the number of data units to be moved by the recently increased latency index. The host system 110 can then return to 405 to determine if there is data to be moved.

Now, assuming that the wait time for the first entry has not expired 415, does the host system 110 use traditional I / O behavior to move the first data track? You can determine whether or not 420. Determining whether to use traditional I / O behavior 420 includes comparing the remaining estimated wait time with the estimated time to copy data through traditional I / O behavior. be able to. If the traditional I / O operation takes longer, the host system 110 can return to 405 to determine if there is data to be moved. Assuming that the traditional I / O operation takes less time than the expected remaining latency, the host system 110 moves the first data track through the traditional I / O operation. 435 can be done. After moving the data, the host system 110 can record the first data track as being moved 440 and return to 405 to determine if there is any data to be moved.

In some embodiments, if the host system 110 sends an unsuccessful fast replication request to the data storage subsystem 120 for each entry in the entry list, the host device 110 will be in which entry. It also returns to the top of the list of remaining entries without successful processing, even if the expected wait time for the first entry does not exceed the expected duration for traditional I / O operation. The / O operation can copy the first entry. Under these conditions, by copying the first entry through traditional I / O operation, the host system 110 can copy the data while waiting for the expected wait time for the remaining entries to expire. become.

The present invention improves system efficiency with respect to copying or moving data. The present invention can be embodied in other specific forms without departing from its gist or essential features. The embodiments described are merely exemplary in all respects and are considered to be non-limiting.