TWI509426B - System for achieving non-interruptive data reconstruction - Google Patents

Description

System for achieving interference-free data reconstruction

The present invention relates to a system for achieving data reconstruction, and more particularly to a system for a cloud host for achieving interference-free data reconstruction. When the system is running, the host can operate smoothly, and the accessed data has been copied or moved to another storage device by a storage device.

For cloud services, there are usually several important components, such as an application server (host) and many storage devices to operate. The storage device may include some hard disks and form a disk array, some logical unit number devices are assigned to the disk array, or the storage device may include some hard disks and store a lot of object storage materials. In order to prevent the storage device from aging and partial damage, upgrade to a faster storage device for better performance, or downgrade to multiple more cost-effective storage devices for cost reasons, after a period of use, a storage device The stored data in the (or logical unit number device) should be copied or moved to another storage device (or logical unit number device). Data reconstruction can be a method of data backup and data migration. The storage devices can be placed in different locations. Data transmission can be achieved via a regional or wide area network.

Traditionally, there are three methods for data copying or moving (rebuilding), namely offline methods, mirroring methods, and double-write and snapshot image copying methods. For the offline method, please refer to Figure 1. The three steps are applied to the offline method. Step 1. An application server 1 accesses data in a source storage device 2, and the access operation is proposed by the user. In step 2, a data replication server 4 is responsible for data reconstruction from the source storage device 2 to a target storage device 3. At the same time, the application server 1 is turned off and no service is provided. When the data reconstruction in the target storage device 3 is completed in step 3, the application server 1 is powered on and connected to the target storage device 3. Thus, the service provided by the application server 1 can continue. Although the offline method can easily implement data reconstruction, it has the problem of service interruption. It is unacceptable for busy cloud services.

The mirroring method is illustrated in Figure 2 and also contains three main steps. Step 1 describes an application server 1 linking and accessing a source storage device 2, and the data in the source storage device 2 is about to be reconstructed in a target storage device 3. In step 2, the source storage device 2 starts mirroring the service and copies the data it contains into the target storage device 3. After the completion of step 2, the source storage device 2 exchanges roles with the target storage device 3, and the application server 1 connects and accesses the target storage device 3 (now it becomes the source storage device) as shown in step 3 of FIG.

Unlike offline methods, the storage devices used in the mirroring method must be provided by the same vendor or by a single stored virtual machine monitor. (Hypervisor) is provided by the run. The mirroring feature must be available for all storage devices at the same time. If the efficiency of the target storage device 3 is inferior to that of the source storage device 2, the application server 1 will face a service interruption. At the same time, additional storage device image synchronization will affect the performance of the source storage device 2. Most importantly, the application server 1 needs to be rebooted to switch the identification characteristics of the source storage device to the identification characteristics of the target storage device, which in turn causes another interference.

The double write and snapshot image copy methods are illustrated in Figure 3. An external in-band data mobile service unit 9 is coupled to an application server 8, a source storage device 10, and a target storage device 11. A planned downtime is required for implementing the intra-frequency infrastructure of the data movement service unit 9. The data movement service unit 9 issues a snapshot command to the source storage device 10 and begins copying the snapshot image to the target storage device 11. The "double write" is because the data movement service unit 9 executes all write commands to the source storage device 10 and the target storage device 11 after the snapshot time point. Finally, when the snapshot image copy is complete and all new writes are completed, the source storage device 10 exchanges roles with the target storage device 11.

In addition to the bottleneck of service interruption or interruption when the intra-frequency infrastructure causes data reconstruction, the source storage device 10 needs to perform both reading and writing operations for a period of time until both snapshot image copying and new write command operations are completed, double writing And the snapshot image copying method also causes the application server 8 to face a service interruption. Therefore, the double write and snapshot image copy methods are still inconvenient for data copying or moving.

Therefore, a system for achieving interference-free data reconstruction, particularly a storage device for a server serving a cloud service, is highly desirable.

This paragraph of text extracts and compiles certain features of the present invention. Other features will be revealed in subsequent paragraphs. The intention is to cover various modifications and similar arrangements in the spirit and scope of the appended claims.

As noted above, existing systems for data movement or replication are disruptive and therefore require a system for achieving interference-free data reconstruction. According to one aspect of the present invention, a system for achieving interference-free data reconstruction includes: a source storage device including a source material; a target storage device; and a server for accessing the source storage device The source data and any data stored in the target storage device are installed with a proxy module and a path input and output module, wherein the proxy module continuously collects a performance parameter value, and the performance parameter value expresses the server source data. Accessing the status, and after the source storage device snapshot, recording the changed block in the source storage device into a meta-chart generated by the path input and output module; the path input and output module will The storage device name of the source storage device connected to the server is decoupled and coupled to the storage device name of the target storage device to achieve the server connection path, switching from the source storage device to the target storage device; a traffic model Unit, receiving the performance parameter value, and according to the received performance parameter value, at a specific time point in the future A data access traffic state to define a strategy in accordance with the state to access traffic information, provided An action schedule; and at least one data movement service unit for snapshotting a portion or the entire source storage device, copying a snapshot of the portion or the entire source storage device to the target storage device, and comparing the action schedule In the previous snapshot, the change in a new snapshot of the part or the entire source storage device is modified to the copied snapshot in the target storage device and controls the operation of the proxy module.

The proxy module communicates with the traffic model unit and at least one data movement service unit; the bit map includes one bit with a value of 0 or 1 for each block in the source storage device; 0 represents no corresponding block Change and 1 represents a change in the content of the corresponding block between two consecutive snapshots; when the number of 1 in the bit chart is less than or equal to a threshold, switching is performed.

According to the present invention, the source storage device is a hard disk, a solid state hard disk, a random access memory, or a mixture of the foregoing devices, or a logical unit number device, or an object storage material. Similarly, the target storage device is a hard disk, a solid state hard disk, a random access memory, or a mixture of the foregoing devices, or a logical unit number device, or an object storage material.

The proxy module is software, hardware or firmware. Similarly, the path input/output module is software, hardware or firmware.

According to the invention, the performance parameter is the number of input and output operations per second, the delay time, and/or the throughput. The action schedule specifies the time to snapshot the portion or the entire source storage device, copying a snapshot of the portion or the entire source storage device to the target storage device, compared to the previous snapshot, the portion or The change in a new snapshot of the entire source storage device is modified to the copied snapshot in the target storage device, and the switch is performed when the source storage device does not have or has a small amount of data access. The small amount of data access is defined in a service level protocol, a quality of service requirement, or a defined policy based on the source storage device or target storage device specification.

The definition policy defines a lag time, and once the action fails to meet the requirements in a service level agreement or a service quality requirement, the copying of the snapshot of the part or the entire source storage device to the target storage device is stopped, and will be compared to the previous A snapshot, a change in a new snapshot of the portion or the entire source storage device, stops modifying the copied snapshot to the target storage device. The definition policy may also define a wait time, but when the action fails to meet the requirements in a service level agreement or a service quality requirement, retry copying the snapshot of the part or the entire source storage device to the target storage device, or A modified snapshot to a copy of the target storage device is compared to the previous snapshot, a change in a new snapshot of the portion or the entire source storage device. The storage device name includes the model name, model number, and/or serial number of the source storage device or the target storage device.

In order to operate the system, a program includes the steps of: A. the traffic model unit provides the action schedule; B. the data move service unit indicates a portion of the source storage device snapshot or the entire source storage device, and the proxy module records the source The changed block in the storage device is in the bit chart, and the bit chart is initially 0; C. The data mobile service unit is in accordance with the action schedule. Copying the snapshot to the same block in the target storage device in sequence; D. the data mobile service unit modifies the change of the partial or entire source storage device snapshot to the copied snapshot in the target storage device; E. repeat step B Go to step D until the number of 1 in the bitmap is less than or equal to the threshold; and F. the data movement service unit performs the switching of the server connection path from the source storage device to the target storage device, and/or Reconnect to the source storage device as required by the user. Here, the threshold is 0.

The above procedure further includes a step B1 between step B and step C, and a step G after step F: B1. If step B to step D have been repeated at least once, the number of 1 in the bit chart is less than the threshold value. And the data access traffic status is expected to be low for a period of time in the near future, so that the data mobile service unit calculates that the copy of the snapshot can be completed within a time required by a service level protocol, and the data mobile service unit starts The proxy module retains all data write commands to the source storage device; and G. the data move service unit instructs the proxy module to release all data to the target storage device and/or the source storage device.

The program can further include a step B2 between step B and step C, and a step E1 between step E and step F: B2. If step B to step D have been repeated at least once, the number of 1 in the bit map is less than The threshold value, and the data access traffic status is expected to be low for a period of time in the near future, so that the data movement service unit calculates that the copy of the snapshot can be completed within a time required by a service level protocol, the proxy mode The group informs the data mobile service unit to close And a new block write command to the source storage device; and E1. the data move service unit executes the new block write command to a corresponding block in the target storage device. The threshold can be any integer less than 10.

With the help of the traffic model unit, it is possible to predict a low amount of access demand for a period of time, and data reconstruction (copying or moving) can be performed during that time. Therefore, data reconstruction can be completed without any interference or interruption of the services provided by the system, that is, to achieve non-interfering data reconstruction. At the same time, the system can be applied regardless of whether the source storage device or the target storage device is from the same or different manufacturer, or the storage device is a physical device or a logical unit number device, or even an object storage material.

1‧‧‧Application Server

2‧‧‧Source storage equipment

3‧‧‧Target storage equipment

4‧‧‧Data Copy Server

8‧‧‧Application Server

9‧‧‧Data Mobile Service Unit

10‧‧‧Source storage equipment

11‧‧‧Target storage equipment

100‧‧‧ system

110‧‧‧Source storage equipment

120‧‧‧Target storage equipment

130‧‧‧Server

140‧‧‧Flow model unit

150‧‧‧Data Mobile Service Unit

151‧‧‧Auxiliary Data Mobile Service Unit

Figure 1 shows a system that operates by an offline method.

Figure 2 shows a system that operates by a mirroring method.

Figure 3 shows a system that operates by a pair of write and snapshot image copy methods.

Figure 4 shows a system for achieving interference-free data reconstruction in accordance with the present invention.

Figure 5 illustrates a one-dimensional chart used in the present invention.

Figure 6 is a flow chart of the procedure for operating the system.

Figure 7 shows another system for achieving interference-free data reconstruction in accordance with the present invention.

Figure 8 illustrates how data is copied between a source storage device and a target storage device.

The invention will be more specifically described by reference to the following examples.

Referring to Figures 4 through 6, an embodiment of the present invention is disclosed herein. A system 100 that achieves interference-free data reconstruction is shown in FIG. The term "reconstruction" means that data in a storage device can be copied or moved to another storage device. If the data is moved to another storage device, the data in the original source storage device will not be accessed or used after the rebuild is completed. Data reconstruction can be done without interference so that one or more servers accessing the original storage device will not be interrupted, and the services provided may remain unchanged. According to the present invention, the system 100 includes at least one source storage device 110, a target storage device 120, a server 130, a traffic model unit 140, and a data movement service unit 150. The function and operation of the above device will be explained in detail below.

The source storage device 110 and the target storage device 120 used in this embodiment are both hard disks. The characteristics of the source storage device 110 and the target storage device 120, such as capacity, rotational speed, power consumption, etc., may be the same, partially identical, or completely different. That is to say, the two storage devices can be from the same or different manufacturers, unlike the mirroring method, which requires two storage devices to have consistent specifications. In practice, the source storage device 110 or the target storage device 120 may be a solid state hard disk, a random access memory, or a hard disk, a solid state hard disk, and a random storage. Take a mixture of at least two of the memory. The source storage device 110 or the target storage device 120 can also be a logical unit number device, or an object storage material for data storage purposes.

The number of physical storage devices in the source storage device 110 and the target storage device 120 is not limited to one. For example, in another embodiment, the source storage device 110 can include a plurality of hard disks that form a disk array, and the target storage device 120 is still a hard disk having storage equivalent to the disk array. Equipment capacity. The source storage device 110 can contain a number of materials or files waiting to be copied or moved. For a complete understanding of the present invention, only one source of material is stored in source storage device 110 for illustration.

The server 130 is used to access source material in the source storage device 110. Therefore, the server 130 can provide a specific service, such as a music stream. The server 130 can also access any of the data 120 stored in the target storage device. However, it should be emphasized that in this embodiment, the connection between the server 130 and the source storage device 110, or the connection between the server 130 and the target storage device 120 may be directly connected, connected via a regional network, or via a wide area network. , such as the Internet, connected. Server 130 can provide this service to a limited group of users on a regional or wide area network. The server 130 has a proxy module and a path input and output module (not shown). The proxy module or path input/output module can exist in a soft, hardware or firmware form to perform its functions. In this embodiment, the proxy module and the path input/output module are both software installed on the server 130.

Regarding the function, the proxy module can continuously collect a performance parameter value that expresses the access status of the server 130 to the source material. Here, the performance parameter is the number of input and output operations per second. The historical data accessed by these users will be sent to the traffic model unit 140 via the server 130 for further analysis. There are other available performance parameters for data access analysis, for example, delay time and/or throughput can be used. As long as the value can be accepted by the traffic model unit 140, the present invention does not limit the collection of specific performance parameter values, as will be described in detail later. The proxy module can record the changed block in the source storage device 110 to a one-bit graph after the snapshot of the source storage device 110. The bitmap is generated by the path input and output module. Because system 100 is used for data copying and movement, snapshots are primarily used to achieve this goal, and the bitmaps used by the proxy modules are stored in server 130. How to track and control the difference between the source storage device 110 snapshot and the target storage device 120 copy snapshot, because the write command after the snapshot is important.

Please refer to Figure 5. For each block in source storage device 110, the bit map contains a single bit with a value of 0 or 1. 0 means that there is no change in the corresponding block and 1 means that the content in the corresponding block between two consecutive snapshots has changed. For example, when a snapshot occurs at T ₀ , the value of all corresponding 24 blocks in the bitmap array is ₀ because there is no change in the data structure. In fact, the data structure contains only source material. The source data has not been overwritten, so the snapshot remains unchanged. When the time comes to T ₁ , another snapshot is taken. At this point, the source data has been overwritten and some of the blocks have changed. There are 4 changed blocks, and the changed parts are C ₀ → C ₁ , R ₀ → R ₁ , S ₀ → S ₁ , and Z ₀ → Z ₁ . The corresponding bit in the array becomes 1 while the other bits are still 0. When the time comes to T ₂ , the source storage device takes a third snapshot. Source data is overwritten again and some blocks are still being changed. The change between T ₁ and T ₂ is F ₀ → F ₂ and R ₁ → R ₂ . It is obvious that the block R ₂ changes twice and the block F ₂ changes only once. However, no matter how many blocks change due to a write command, the corresponding bit will change from 0 to 1. Finally, at _3:00 T, there is no change between successive snapshots, all the bits are bit charts resumed 0. This is how the bit chart works.

The path input and output module decouples the storage device name of the source storage device 110 connected to the server 130 and is coupled to the storage device name of the target storage device 120. The path input and output module operates in such a manner as to achieve a switch of the server 130 connection path from the source storage device 110 to the target storage device 120. The path input and output module controls the path accessed by the server 130. The storage device name includes the model name, model number, and/or serial number of the source storage device 110 or the target storage device 120, which can contain other information regarding the specifications of the source storage device 110 or the target storage device 120. When the path input and output module decouples the storage device name of the source storage device 110, the server 130 cannot connect to and access the source storage device 110. When the path input and output module is coupled to the storage device name of the target storage device 120, the server 130 is connected to the target storage device 120, and access between the two is feasible. After the action of decoupling and coupling, the path input/output module switches the access path of the server 130.

The traffic model unit 140 can receive the value of the number of input/output operations per second (performance parameter) from the proxy module, and can input and output according to the received input per second. The value of the operation times provides a data access flow status at a specific time in the future. The traffic status is an entity material that confirms how the server 130 accesses the source storage device 110. For example, the traffic status can be the number of input/output operations per second at any time after 5 minutes. However, the traffic status must be gathered from the number of input/output operations per second accumulated in past history and analyzed. Any suitable method, algorithm or module for providing such a service can be applied, preferably using a stored flow model provided by the same inventor in U.S. Patent Application Serial No. 14/290,533, which is incorporated herein by reference. A common reference for the same technology. Therefore, the traffic model unit 140 can provide an action schedule according to a defined policy and data access traffic status.

Here, the action schedule is a set of instructions for operating the system 100 at the optimal point in time. For example, the action schedule specifies the snapshot of the portion or the entire source storage device 110, copying the snapshot of the portion or the entire source storage device 110 to the target storage device 120, as compared to the previous snapshot, the portion or the entire source storage device 110 The change in a new snapshot, the copied snapshot to the target storage device 120, and the time at which the switch is performed when the source storage device 120 does not have or accesses a small amount of data. The foregoing small amount of data access is defined in a service level protocol, a quality of service requirement, or in a defined policy based on the source storage device 110 or the target storage device 120 specification. It is important to emphasize that the entire action schedule is based on an estimate of future traffic status, and the action schedule is modified by the defined strategy and evaluated when an unexpected condition occurs.

The definition policy is different from the service level agreement or service quality requirement, and the latter two are signed and agreed by the manufacturer of the system 100 and the service provider owning the system 100 before the system 100 goes online. The definition policy sets up the operation of the detailed system 100 to implement service content for service level agreements or service quality requirements. In particular, the definition policy defines a lag time, and once the action fails to meet the requirements in a service level agreement or a service quality requirement, the copying of the snapshot of the part or the entire source storage device 110 to the target storage device 120 is stopped, and the phase is The change to a new snapshot of the portion or of the entire source storage device 110, compared to the previous snapshot, stops modifying the copied snapshot into the target storage device 120. For example, if the service level agreement requires the server 130 to respond to a user request for no more than 2 seconds, when the snapshot copy time exceeds 5 seconds of lag time, and in this case, the response time may be longer than 2 seconds, then The entire copy operation stops. The copy of the snapshot will be delayed by 30 seconds, or until the access expected by the server 130 is lower than the number of input/output operations per second, such as 10,000 Mb/s, which does not delay the operation of the snapshot copy. In another example, if there are some blocks between the two snapshots that need to be updated by some receive write commands, the definition policy will require the shortest time, such as 3 seconds, to process the action to avoid service level agreements or quality of service requirements. The requirements in can not be met. The "5 second" lag time, the "30 second" wait, and the "3 second" minimum time are some aspects of defining the strategy. The snapshot of the source storage device 110 may be many times before the data copy or mobile service is completed, and the definition policy is applicable for each time.

As can be seen from the above description, the definition policy also defines a waiting time. Once the action cannot meet the requirements in a service level agreement or a service quality requirement, retry copying the snapshot of the part or the entire source storage device 110 to the target storage device. 120. Depending on the previous snapshot, the change in a new snapshot of the portion or the entire source storage device 110, the modified snapshot to the target storage device 120 is retried. For example, a "30 second" wait.

The data movement service unit 150 can access the source storage device 110 and the target storage device 120, and communicate with the server 130 and the traffic model unit 140, so it is responsible for snapshotting a portion or the entire source storage device 110, a snapshot portion or the entire source storage device 110. Copying a snapshot of the portion or the entire source storage device 110 to the target storage device 120, and in accordance with the action schedule, a change in a new snapshot of the portion or the entire source storage device 110 compared to the previous snapshot, The copied snapshot to the target storage device 120 is modified. It also controls the operation of the proxy module. It should be noted that the control function can be completed because the proxy module can communicate with the traffic model unit 140 and the data movement service unit 150. From another point of view, the proxy module is a window of the server 130 for receiving control commands from the data movement service unit 150. For load distribution and balancing, there may be more than one data movement service unit 150 in the system, as will be described in detail in another embodiment that follows.

The program of the operating system 100 will now be described as follows. The program consists of several steps, please refer to Figure 6, which is a flow chart of the program. Before the data movement, the traffic model unit 140 continuously collects the value of the number of input and output operations per second. This value expresses the access status of the source data by the server 130. Based on the collected values, the traffic model unit 140 can provide an action schedule (S01). When the data movement service starts, the data movement service unit 150 instructs the source storage device 110 to snapshot the entire source storage device 110, which also instructs the agent module to record the changed blocks in the source storage device 110 to the bits in the server 130. Chart (S02). As described above, since the block change is caused by the write command to the source storage device 110 between the two snapshots, all the bits contained in the bit map are 0 to correspond to all the blocks in the initial stage. When the data movement service unit 150 confirms that a snapshot has been generated, it copies the snapshot to the same block in the target storage device 120 in order according to the action schedule (S03). In step S03, the material movement service unit 150 will follow the action schedule from the traffic model unit 140 to perform snapshot copying when the source storage device 110 anticipates no access or low access. Thus, the service interruption caused by the access by the server 130 can be avoided. Because the snapshot copying continues, only some of the blocks that have changed due to the write command, the entire replicated snapshot in the target storage device 120 is rarely written. The mobile service unit 150 modifies only the changed portion of the entire source storage device 110 snapshot to the snapshot copied in the target storage device 120 (S04). As shown in Figure 5, the number of changed blocks may be reduced in the next snapshot. By repeating step S02 to step S04, convergence on the snapshot copy time can be achieved. The repetition of these steps will continue until the number of 1s in the bit map is less than or equal to the threshold value (S05). Preferably, the threshold is zero, which means that there is no block change between successive snapshots of the source storage device 110. Next, the data movement service unit 150 performs switching of the connection path of the server 130. From the source storage device 110 to the target storage device 120, or reconnected to the source storage device 110 (S06), depending on the needs of the user.

In detail, step S06 includes a sub-step of switching that stops the output input between the server 130 and the source storage device 110. After the handover, the data movement service unit 150 notifies the path input/output module via the proxy module, changes the connection path of the server 130, and becomes the target storage device 120 by the source storage device 110. Thus, the target storage device 120 becomes a new source storage device. Alternatively, the data movement service unit 150 will notify the path input/output module via the proxy module to reconnect the path of the server 130 to the source storage device 110, depending on the needs of the user. As for the timing of the handover, the same is true for the source storage device 110 and the target storage device 120. In the data copying process, when the access amount is large, because there is the traffic model unit 140, the server 130 does not need to perform copy snapshot and switching, and the service provided by the server 130 is not affected, and the interference-free data reconstruction can be achieved (replication or mobile).

For the above description, it is obvious that the method for switching is to wait for no block change between two consecutive snapshots. As is known, the fact that two consecutive snapshots have no write commands is ideal. Therefore, a more practical switching method or operating procedure is needed, and the present invention provides other solutions.

If step S02 to step S04 have been executed at least once, the number of 1 in the bit map is less than the threshold, and the data access flow state is expected to be low for a period of time in the near future, so that the data movement service unit 150 can calculate Know that the copy of the snapshot can be completed within the time required by a service level agreement. The mobile service unit 150 activates the proxy module to retain all data write commands to the source storage device 110. Depending on the needs of the user, the data movement service unit 150 instructs the agent module to release all data write commands to the target storage device 120 or the source storage device 110 after step S06. Preferably, the threshold is any integer less than 10, such as 5. All data is written in sequence before switching.

If the number of 1 in the bit chart is less than the threshold value, the number of 1 in the bit chart is expected to be low for a period of time in the near future, so that the data movement service is available, in the same premise that step S02 to step S04 have been repeated at least once. The unit 150 can calculate that the copy of the snapshot can be completed within a time required by a service level protocol, and the proxy module can notify the data movement service unit 150 of a new block write command to the source storage device 110. After the block modification changed between the two consecutive snapshots is completed, the material movement service unit 150 performs a new block write command to the corresponding block in the target storage device 120. Next, step S06 is performed. Similarly, the threshold is any integer less than 10. Thus, all data writes are completed in sequence before switching.

In addition, if the service level agreement requires a shorter response time or the service provider wants to speed up the operation of the snapshot copy, more than one data move service unit can be used. See Figure 7. An auxiliary material mobile service unit 151 is connected to the system 100. As noted above, the snapshot can be completed for a portion of the source storage device 110, rather than the entire source storage device 110. Please also refer to Figure 8. If the source storage device 110 is divided into two parts, the track 1 to the track 100 and the track 101 Go to track 200. The target storage device 120 is also divided into two parts, a track 1 to a track 100 and a track 101 to a track 200. The data movement service unit 150 is responsible for copying or moving the source material from the track 1 to the track 100 block of the source storage device 110 to the track 1 to the track 100 of the target storage device 120. The auxiliary material mobile service unit 151 is responsible for copying or moving the source material from the source storage device 110 track 101 to the track 200 block to the track 101 of the target storage device 120 to the track 200. Traditionally, track 1 to track 100 of source storage device 110 is defined as a logical unit number device, and track 101 to track 200 of source storage device 110 is defined as another logical unit number device. Thus, the application of the present invention is not limited to data copying or data movement between two physical storage devices. Logical storage devices can also be applied.

According to the above embodiment, it should be noted that if the server 130 has some functions of the modules before the proxy module and the path input/output module are installed, these functions should be retained while other servers 130 do not. The function is implemented by the proxy module and the path input and output module. For example, the path input and output module can utilize the existing path input and output elements in the server 130 operating system, which provides only the modification and insertion functions required for data movement or copy services. Similarly, if the server 130 or source has a similar copy service to the operating system of the target storage device, the data movement service unit 150 will utilize the data copy service on the local machine. In addition, based on the defined policy set by the service provider, a positive snapshot copy or a final switch can be achieved during the expected low access time. In other words, defining a strategy is a way of controlling how efficiently data replication or movement can be done. The data movement service unit 150 can further obtain when the access amount is low A snapshot copy speed reference between the source storage device 110 and the target storage device 120. The baseline can be sent to the traffic model unit 140 to create an action schedule and calculate the expected completion time for each snapshot copy.

Although the present invention has been disclosed in the above embodiments, it is not intended to limit the invention, and those skilled in the art can make some modifications and refinements without departing from the spirit and scope of the invention. The scope of the invention is defined by the scope of the appended claims.

100‧‧‧ system

110‧‧‧Source storage equipment

120‧‧‧Target storage equipment

130‧‧‧Server

140‧‧‧Flow model unit

150‧‧‧Data Mobile Service Unit

Claims (17)

A system for achieving non-interfering data reconstruction, comprising: a source storage device, including a source data; a target storage device; and a server for accessing source data stored in the source storage device and stored in the source Any data in the target storage device, a proxy module and a path input and output module are installed, wherein the proxy module continuously collects a performance parameter value, the performance parameter value expresses an access status of the server to the source data, and After taking a snapshot of the source storage device, the changed block in the source storage device is recorded into a one-dimensional chart, and the bit chart is generated by the path input and output module; the path input and output module is connected to the The storage device name of the source storage device of the server is decoupled and coupled to the storage device name of the target storage device, so as to achieve the switching of the server from the source storage device to the target storage device connection path; a traffic model unit receives the Performance parameter value, and can provide a data access at a specific time point in the future according to the received performance parameter value a quantity status to provide an action schedule in accordance with a defined policy and the data access status; and at least one data movement service unit for taking snapshots of the source material in a portion or the entire source storage device The source data in the copy is copied to the target storage device, and according to the action schedule, Re-snapping the source material in the part or the entire source storage device. If there is a difference between the re-taken snapshot and the previous snapshot, the difference between the two is modified to the previous snapshot. Copying source data to the target storage device and controlling operation of the proxy module; wherein the proxy module communicates with the traffic model unit and at least one data movement service unit; the bit map is for each of the source storage devices A block contains a bit with a value of 0 or 1; 0 means no change in the corresponding block and 1 means change in the corresponding block between two consecutive snapshots; when the number of 1 in the bit chart is less than or equal to one When the threshold is exceeded, a switch is performed. The system of claim 1, wherein the source storage device is a hard disk, a solid state hard disk, a random access memory, or a mixture of the foregoing devices, or a logical unit number device, or an object storage material. . The system of claim 1, wherein the target storage device is a hard disk, a solid state hard disk, a random access memory, or a mixture of the foregoing devices, or a logical unit number device, or an object storage data. . The system of claim 1, wherein the proxy module is a soft body, a hardware or a firmware. The system of claim 1, wherein the path input/output module is a software, a hardware or a firmware. The system of claim 1, wherein the performance parameter is an input/output operation number per second, a delay time, and/or a throughput. The system of claim 1, wherein the action schedule specifies a time at which the source material in the portion or the entire source storage device is snapshotted, and the source data in the snapshot range is copied to the target storage device. Time, the time to re-snap the source material in the part or the entire source storage device, compare the retaked snapshot with the previously taken snapshot, and modify the difference between the two to the previous snapshot The time at which the source data is copied to the target storage device and when the source storage device is expected to have no data access or data access is small. The system of claim 7, wherein the small amount of data access is defined in a service level agreement, a quality of service requirement, or a defined policy based on the source storage device or the target storage device specification. The system of claim 1, wherein the definition policy defines a lag time, and if the action fails to meet a requirement in a service level agreement or a service quality requirement, stopping the part or the entire source within the snapshot range The source data in the storage device is copied to the target storage device, and the difference between the retaken snapshot and the previous snapshot is stopped to be changed to the source data copied to the target storage device in the previous snapshot. The system of claim 1, wherein the definition policy defines a waiting time, and if the action fails to meet a requirement in a service level agreement or a service quality requirement, retrying the part or the whole of the snapshot range Copy the source material from the source storage device to the target storage device, or retry to modify the difference between the retaken snapshot and the previous snapshot to the source copied to the target storage device from the previous snapshot. data. The system of claim 1, wherein the storage device name includes a model name, a model number, and/or a serial number of the source storage device or the target storage device. The system of claim 1, wherein the program for operating the system comprises the steps of: A. the traffic model unit provides the action schedule; B. the data move service unit indicates a portion of the source storage device snapshot Or the entire source storage device, and the agent module records the changed block in the source storage device into the bit chart, the bit chart is initially 0; C. the data mobile service unit is in accordance with the action schedule Copying the snapshot to the same block in the target storage device in sequence; D. the data mobile service unit modifies the change of the partial or entire source storage device snapshot to the copied snapshot in the target storage device; E. repeating the steps B to step D until the number of 1 in the bit map is less than or equal to the threshold value; and F. the data movement service unit from the source storage device to the target storage device performs switching of the server connection path, and/ Or re Connect to the source storage device. The system of claim 12, wherein the threshold is zero. The system of claim 12, wherein the program further comprises a step B1 between step B and step C, and a step G after step F: B1. if step B to step D have been repeated at least once The number of 1 in the bit chart is less than the threshold, and the data access traffic state is expected to be low for a period of time in the near future, so that the data mobile service unit calculates that the copy of the snapshot can be in a service level protocol. After the required time is completed, the data mobile service unit activates the proxy module to retain all data write commands to the source storage device; and G. the data mobile service unit instructs the proxy module to release all data to be written to the Target storage device and/or the source storage device. The system of claim 14, wherein the threshold is any positive integer less than 10. The system of claim 12, wherein the program further comprises a step B2 between step B and step C, and a step E1 between step E and step F: B2. if step B to step D have been Repeating at least once, the number of 1 in the bit chart is less than the threshold, and the data access traffic state is expected to be low for a period of time in the near future, so that the data mobile service unit calculates that the copy of the snapshot can be The time required by the service level agreement is completed, the agent module notifies the data mobile service unit about storing to the source a new block write command for the device; and E1. The data move service unit executes the new block write command to the corresponding block in the target storage device. The system of claim 16, wherein the threshold is any positive integer less than 10.