TWI766409B - Data access system - Google Patents

Abstract

A data access system includes: a data storage medium, a record medium, a first controller, and a second controller. The record medium includes a first record area and a second record area. The first controller includes a first metadata area. The second controller includes a second metadata area. The first controller is connected to the data storage medium and the record medium and corresponds to the first record area. The second controller is connected to the data storage medium and the record medium and corresponds to the second record area. The first controller receives the first data, and writes the first data into the data storage medium in a log manner to update the first metadata area, and correspondingly generates the first record in the first record area. The second controller updates the second meta data area according to the first record in the first record area.

Description

data access system

This case is about the field of storage, especially a data access system.

In recent years, with the rapid development of data access system technology, data access systems have been widely used in many applications in daily life, such as electronic mail servers, financial transactions, database management, and so on. As the data to be processed and the services provided by today's data access systems continue to increase, the operational burden of the data access systems also increases. Therefore, how to reduce the downtime of the data access system during multi-tasking and high-speed processing, that is, to maintain the high availability of the data access system, has become an important issue.

The current data access system with dual controllers is limited by the architectural design of switching access rights between the dual controllers, or limited by the bandwidth of the transmission interface between the dual controllers, making the operation of the data access system impossible. Speed is limited and performance is reduced.

In view of this, this case proposes a data access system.

According to some embodiments, a data access system includes: a data storage medium, a recording medium, a first controller, and a second controller. The recording medium includes a first recording area and a second recording area. The first controller includes a first metadata area. The second controller includes a second metadata area. The first controller is connected to the data storage medium and the recording medium, and corresponds to the first recording area. The second controller is connected to the data storage medium and the recording medium, and corresponds to the second recording area. The first controller receives a first data, and writes the first data into the data storage medium in a log mode to update the first metadata area, and correspondingly generates a first record in the first recording area. The second controller updates the second metadata area according to the first record in the first record area.

According to some embodiments, when the second controller in the aforementioned data access system checks that the occupied capacity of the first record in the first recording area satisfies a threshold, it updates the second element according to the first record in the first recording area. data area.

According to some embodiments, when the first controller in the aforementioned data access system checks that the first recording area is full, the second controller is triggered to update the second metadata area according to the first record in the first recording area, and confirms that After the second controller is updated, the first recording area is cleared.

According to some embodiments, the second controller in the data access system receives a second data, writes the second data into the data storage medium in a log mode to update the second metadata area, and generates a corresponding The second record is in the second record area, and the first controller updates the first metadata area according to the second record in the second record area.

According to some embodiments, when the first controller in the aforementioned data access system checks that the occupied capacity of the second record in the second recording area satisfies a threshold, it updates the first element according to the second record in the second recording area. data area.

According to some embodiments, when the second controller in the data access system checks that the second recording area is full, the first controller is triggered to update the first metadata area according to the second record in the second recording area, and confirms After the first controller is updated, the second recording area is cleared.

According to some embodiments, the aforementioned data access system further includes: a multi-path input and output connected to the first controller and the second controller. The first controller receives the first data through the multi-path input and output, and the second controller receives the second data through the multi-path input and output.

According to some embodiments, the first record in the aforementioned data access system includes: a first time stamp of receiving the first data via multi-path input and output, the first data being stored in a first location on the data storage medium, and the first A first volume of a data. The second record in the aforementioned data access system includes: a second time stamp for receiving the second data via the multi-path input and output, a second location for storing the second data in the data storage medium, and a first time stamp for the second data Two capacity.

According to some embodiments, the aforementioned data access system further includes: a multi-path input and output connected to the first controller. The first controller receives the first data via the multi-path input and output.

According to some embodiments, the first record in the aforementioned data access system includes: a first time stamp of receiving the first data via multi-path input and output, the first data being stored in a first location on the data storage medium, and the first A first volume of a data.

To sum up, the data access system of some embodiments of the present application includes a first controller and a second controller. Through the recording in the recording area, another controller can update the metadata area accordingly, so that the two controllers can synchronize their knowledge of the data in the data storage medium. Therefore, the data access system does not need to be limited by the transmission bandwidth between the dual controllers, nor does it need to rely on the mutual exclusion lock mechanism between the dual controllers, so that the data access system can achieve high-speed access controlled by the dual controllers. Function.

FIG. 1 is a schematic diagram of a data access system 10 according to an embodiment of the present invention. Referring to FIG. 1 , the data access system 10 includes: a connection interface 100 having a multipath I/O (MPIO), a data storage medium 200 , a recording medium 300 , a first controller 400 and a second controller 500 . The recording medium 300 includes a first recording area 310 and a second recording area 320 , the first controller 400 includes a first metadata area 410 , and the second controller 500 includes a second metadata area 510 . The connection interface 100 is a network interface. The connection interface 100 is connected to the first controller 400 and the second controller 500 for the first controller 400 and the second controller 500 to receive commands from the client and operate according to the commands. The first controller 400 and the second controller 500 are connected to the data storage medium 200 and the recording medium 300 respectively, so as to access the content stored in the data storage medium 200 and the recording medium 300 .

The data storage medium 200 and the recording medium 300 are, for example, but not limited to, electronically erasable programmable read-only memory (EEPROM), phase-change memory (PCM), magnetic Resistive random access memory (Magnetoresistive Random Access Memory, MRAM), variable resistive memory (Resistive Random Access Memory, ReRAM), or flash memory (Flash Memory) and other non-volatile memories with access functions Body (Non-volatile Memory, NVM). In other embodiments, the data storage medium 200 and the recording medium 300 are, for example, but not limited to, a Hard Disk Drive (HDD), a Solid-State Drive (SSD), or a Redundant Array of Independent Disks, RAID), etc., wherein the disk array is not limited to being composed of hard disks, solid-state hard disks, or a combination of hard disks and solid-state hard disks. It should be noted that, according to some embodiments, the data storage medium 200 is a disk array whose level is not less than the fourth level disk array (RAID 4). The first metadata area 410 and the second metadata area 510 can be implemented by, for example, but not limited to, a non-volatile memory with an access function, a hard disk, or a solid-state hard disk.

The data access system 10 stores data in the data storage medium 200 through the first controller 400 or/and the second controller 500 . According to the operation mode of the data access system 10, the data are selectively stored in the data storage through the first controller 400 alone, through the second controller 500 alone, and partly through the first controller 400 and partly through the second controller 500. Media 200. Here, the data stored by the first controller 400 is defined as the first data, and the data stored by the second controller 500 is defined as the second data.

In some embodiments, the data access system 10 is suitable for connecting at least one client 20 to execute a write command of the client 20, to store the data provided by the client 20 in the data storage medium 200; or to execute the client 20. The read command of 20 transmits the data in the data storage medium 200 to the client 20 . The client 20 has a connection interface 21 corresponding to the connection interface 100 of the data access system 10 .

The recording medium 300 includes a first recording area 310 and a second recording area 320 . The first recording area 310 is for the first controller 400 to store the records of writing and deleting performed on the data storage medium 200 (hereinafter referred to as the “first record”). The first record is written by the first controller 400 in a log mode. corresponding to the data. Similarly, the second recording area 320 is for the second controller 500 to store the records of writing and deleting performed on the data storage medium 200 , and the second record is correspondingly generated when the second controller 500 writes data in a log mode. The writing of data in the log mode refers to sequentially writing the data to the physical address of the data storage medium 200, and generating a mapping relationship between the physical address and the logical address, and whether the data is valid or invalid ( For example, executing an instruction to delete the data is a valid value for marking the data as invalid. For details, please refer to the applicant's ROC Patent No. I609323.

The first controller 400 internally has a first metadata area 410 for storing the aforementioned records of the mapping relationship between physical addresses and logical addresses and corresponding valid values (hereinafter referred to as “mapping table”). Similarly, the second controller 500 internally has a second metadata area 510 for storing the aforementioned records of the mapping relationship between physical addresses and logical addresses and corresponding valid values (ie, a mapping table).

In some embodiments, the first metadata area 410 and the second metadata area 510 are stored in the data storage medium 200 . In some embodiments, the first metadata area 410 and the second metadata area 510 are stored in the recording medium 300 .

In some embodiments, the data access system 10 is adapted to operate in a dual active mode and a single active mode. The first controller 400 and the second controller 500 are adapted to operate in an active state or a standby state. When the controller (ie, the first controller 400 or the second controller 500) is in the active state, the controller can receive data and write the data in the data storage medium 200, and can check the recording areas corresponding to other controllers to Get the write status of other controllers, and use this to update your own metadata area. For example, when the first controller 400 operates in the active state, it can receive data and write the data into the data storage medium 200 , and can update the first metadata area 410 according to the second recording area 320 corresponding to the second controller 500 . When the controller is in the standby state, the controller cannot write data into the data storage medium 200, but can check the recording area of other controllers to obtain the writing status of other controllers, and update its own metadata area accordingly . For example, when the second controller 500 operates in a standby state, the second metadata area 510 can be updated according to the first recording area 310 corresponding to the first controller 400 . When the data access system 10 operates in the dual-active mode, both the first controller 400 and the second controller 500 operate in the active state. Conversely, when the data access system 10 operates in the single-active mode, one of the first controller 400 and the second controller 500 operates in the active state, and the other of the first controller 400 and the second controller 500 operates in the active state. The operator operates in the standby state.

FIG. 2 is a schematic diagram of a data storage medium 200 according to some embodiments of the present invention. Referring to FIG. 2 , in some embodiments, the data storage medium 200 is a Redundant Array of Independent Disks (RAID), and the data storage medium 200 includes a plurality of physical blocks (PBs) PB1 , PB2 , and PB3 , PB4, multiple stripes (Stripes) S1, S2, S3, S4, but this case is not limited to this, the number of physical blocks and stripes can be increased or decreased according to the actual needs of the data storage medium 200.

The strips S1-S4 span the physical blocks PB1-PB4 respectively. Specifically, the area of the strips S1-S4 spanning the physical blocks PB1-PB4 includes a plurality of pages (Page), wherein the strip S1 corresponds to the physical area The pages of the blocks PB1-PB4 are respectively page 211, page 221, page 231, and page 241, wherein the respective pages of the strip S2 corresponding to the physical blocks PB1-PB4 are page 212, page 222, page 232, and page 242, respectively, Wherein the strip S3 corresponds to the pages of the physical blocks PB1-PB4 are page 213, page 223, page 233, and page 243, and the strip S4 corresponds to the pages of the physical blocks PB1-PB4 are page 214, page 243, respectively 224, page 234, page 244.

Please refer to FIG. 1 and FIG. 2 simultaneously. In some embodiments, pages 211 to 214 , 221 to 224 , 231 to 234 , and 241 to 244 are used for storing data and corresponding error correction codes (Error Control Codes, ECC). In some embodiments, the first controller 400 may be capable of receiving data to be written (ie, one or more first data), and the first controller 400 stores the first data in pages 211 to 214 and 221 to 224, respectively , 231 to 234, 241 to 244. And the first controller 400 stores the error correction codes corresponding to the first data in pages 211 to 214 , 221 to 224 , 231 to 234 , and 241 to 244 . Similarly, in some embodiments, the second controller 500 can receive data to be written (ie, one or more second data), and the second controller 500 stores the second data in pages 211 to 214 and 221 respectively to 224, 231 to 234, 241 to 244. And the second controller 500 stores the error correction codes corresponding to the second data in pages 211 to 214 , 221 to 224 , 231 to 234 , and 241 to 244 .

In some embodiments, when any one of the physical blocks PB1 to PB4 is faulty, the data storage medium 200 can perform a parity check (Parity Check) through the error correction code to detect errors, and the data storage medium 200 further restores the data stored in the faulty physical block through the data and the error correction code of the other non-faulty physical blocks in the same stripe.

It should be noted that, according to some embodiments, the first controller 400 and the second controller 500 will not write to the same stripe at the same time, for example, when the first controller 400 is storing the first data in the stripe S1 , the second controller 500 will not store the second data in the strip S1. Therefore, the first controller 400 and the second controller 500 do not conflict during execution. In some embodiments, the data access system 10 can control the location where the first controller 400 and the second controller 500 write to the data storage medium 200 , thereby preventing the first controller 400 and the second controller 500 from writing the same location at the same time. a strip. According to some embodiments, the data access system 10 prevents the first controller 400 and the second controller 500 from simultaneously writing to the same stripe through an instruction from a coordination controller (not shown). The coordination controller is connected to the first controller 400 and the second controller 500 to transmit coordination commands to the first controller 400 and the second controller 500, and distribute the first controller 400 and the second controller 500 through the coordination command The respective written stripes. In some embodiments, it is the first controller 400 or the second controller 500 that allocates the stripes that it and the other controller respectively write to.

FIG. 3 is a schematic diagram of a recording medium 300 according to some embodiments of the present invention. Referring to FIG. 1 and FIG. 3 , in some embodiments, the first controller 400 operating in the active state writes the first data into the data storage medium 200 in a log mode to update the first metadata area 410 , and generates a corresponding A first record R1 is in the first record area 310 . That is, when the first controller 400 writes the first data in the log mode, it writes the first data to the physical address of the data storage medium 200 in sequence, and records information about the physical address and the physical address in the mapping table. The relationship between the logical address and the corresponding valid value. In addition, a first record R1 is created according to the changed content, and the content of the first record R1 will be described later. In order to know the latest state of the content stored in the data storage medium 200 , the second controller 500 updates the second metadata area 510 according to the first record R1 in the first record area 310 . The “update” is to add the mapping relationship and valid value (ie, the mapping table) corresponding to the operations of writing, deleting, etc. performed by the first controller 400 to the second metadata area 510 . In this way, the second controller 500 does not need to use an additional transmission channel connected to the first controller 400 or rely on the mutual exclusion mechanism between the two controllers, and the second controller 500 can obtain the first controller 400 related write information. Similarly, for actions such as writing and deleting performed by the second controller 500 , the first controller 400 can also update the first metadata area 410 via the second record R2 in the second record area 320 .

Please continue to refer to FIG. 3 , in some embodiments, the first record R1 includes: a first timestamp T1 , a first location L1 , and a first capacity C1 . The first time stamp T1 is the time stamp at which the connection interface 100 with multi-path input and output receives the first data. The first location L1 is, for example, but not limited to, the physical address, the logical address, or the combination of the physical address and the logical address of the page where the first data is stored in the data storage medium 200 . The first capacity C1 is the data size of the first data. The connection interface 100 generates a corresponding first time stamp T1 according to the time point when the first data is transmitted through the connection interface 100 with multi-path input and output. According to other embodiments, the first time stamp T1 is a time point when the first data is output to the first controller 400 via the connection interface 100 . Similarly, in some embodiments, the second record R2 includes a second timestamp T2, a second location L2, and a second volume C2. The second time stamp T2 is the time stamp at which the connection interface 100 with multi-path input and output receives the second data. The second location L2 is, for example, but not limited to, the physical address, the logical address, or the combination of the physical address and the logical address of the page where the second data is stored in the data storage medium 200 . The second capacity C2 is the data size of the second data. According to other embodiments, the second time stamp T2 is a time point when the second data is output to the second controller 500 via the connection interface 100 .

In some embodiments, a controller (eg, the second controller 500 ) checks the occupied capacity of the first record R1 in the recording area (here, the first recording area 310 ) of another controller (eg, the first controller 400 ) When a threshold value (eg, a specific capacity) is satisfied, the second metadata area 510 is updated (here, it is updated according to the first record R1 of the first record area 310 ). That is to say, the controller checks the recording area of another controller, and only updates the data when the content of the data changes reaches a certain level. In some embodiments, the threshold corresponding to the occupied capacity of the first record R1 is different from the threshold corresponding to the occupied capacity of the second record R2.

In some embodiments, a controller (eg, the first controller 400 ) checks whether its own recording area (here, the first recording area 310 ) is full, and if it is full, triggers another controller (here, the second record area 310 ) The controller 500) updates according to the first record R1 in the first record area 310 (here, the second metadata area 510 is updated). And the controller (here, the first controller 400 ) confirms that the update is completed (here, the second metadata area 510 is updated), and then clears the records in the recording area (here, the first record R1 ).

Table 1 is an example of the data storage medium 200 storing data in some embodiments of the present invention. Table 2 is an example of storing records in the first recording area 310 in some embodiments of the present invention. Table 3 is an example in which no records are stored in the second recording area 320 in some embodiments of the present invention. Table 4 is an example of how the first controller 400 and the second controller 500 determine the data storage sequence in the data storage medium 200 according to the records in the recording medium 300 in some embodiments of the present invention. Tables 1 to 4 are as follows:

Table 1: Data storage medium 200 Physical block PB1 Physical block PB2 Physical block PB3 Physical block PB4 Physical block PB5 Physical block PB6 Physical block PB7 Physical block PB8 Strip S1 D1 D2 D3 D4 D5 D6 D7 P1 Strip S2 D3 D8 D9 D5 D10 D4 D7 P2 Stripe S3 Stripe S4

Table 2: first recording area 310 first timestamp T1 first position L1 first capacity C1 1 (S1, PB1) d1 2 (S1, PB2) d2 3 (S1, PB3) d3 4 (S1, PB4) d4 5 (S1, PB5) d5 6 (S1, PB6) d6 7 (S1, PB7) d7 8 (S2, PB1) d3 9 (S2, PB2) d8 10 (S2, PB3) d9 11 (S2, PB4) d5 12 (S2, PB5) d10 13 (S2, PB6) d4 14 (S2, PB7) d7 . . . . . . . . .

table 3: The second recording area 320 second time stamp T2 second position L2 The second capacity C2 . . . . . . . . .

Table 4: Data storage medium 200 Physical block PB1 Physical block PB2 Physical block PB3 Physical block PB4 Physical block PB5 Physical block PB6 Physical block PB7 Physical block PB8 Strip S1 D1 D2 D3 - old D4 - old D5 - old D6 D7 - old P1 Strip S2 D3-new D8 D9 D5-new D10 D4-new D7-new P2 Stripe S3 Stripe S4

Referring to FIG. 1 and Tables 1 to 4, in some embodiments, the data storage medium 200 is a fourth-level disk array, the stripes include four stripes S1 to S4, and the physical blocks include eight groups of entities PB1 to PB8 block, and the physical block PB8 is used to store the error correction code. It is assumed that neither the data storage medium 200 nor the recording medium 300 originally stored data or records. The first controller 400 operates in an active state. The second controller 500 is not limited to operate in an active state or a standby state.

Continuing from the above, the data access system 10 sequentially writes fourteen pieces of first data through the first controller 400 (respectively D1, D2, D3, D4, D5, D6, D7, D3, D8, D9 according to the writing order) , D5, D10, D4, D7) in the data storage medium 200, wherein the first seven first data are written into the strip S1, and the last seven first data are written into the strip S2. Therefore, the first data stored in the data storage medium 200 is shown in Table 1, wherein the error correction code P1 and the error correction code P2 are the error correction codes corresponding to the stripe S1 and the stripe S2, respectively.

The first record R1 correspondingly generated according to the first controller 400 writing the first data into the data storage medium 200 in a log-like manner is shown in Table 2. The first recording area 310 stores a first time stamp T1, a first position L1 and a first capacity C1. The first time stamp T1 is described by taking “1 to 14” as an example, and the larger the value of the first time stamp T1 represents the newer data. The first position L1 is illustrated by taking "(S1, PB1)" as an example, and represents the page corresponding to the stripe S1 and the physical block PB1. The first capacity C1 is described by taking "d1" as an example, which stores the data size of the first data D1.

Moreover, since the second controller 500 does not write data into the data storage medium 200 , the second record R2 is shown in Table 3, and no record is stored in the second recording area 320 .

When the second controller 500 reads the first recording area 310, the second controller 500 can obtain the relevant information that the first controller 400 writes the first data to the data storage medium 200 in a log-like manner (as shown in Table 2). ), and can determine the data storage sequence in the data storage medium 200 (as shown in Table 4). Herein, "D3-old" and "D3-new" are taken as examples for description, "D3-old" and "D3-new" are respectively stored in different pages in the data storage medium 200, but "D3-old" and "D3- "New" is the data "D3" stored in the data storage medium 200 at different time points, "D3-Old" represents the data stored at an earlier time point, and "D3-New" represents the data stored at a later time point. In some embodiments, data stored at an earlier time point is invalid data and can be erased as junk data. The data stored at a later time point is the valid data, that is, the valid data when read by the data access system 10 .

Furthermore, Table 1 to Table 4 are used as examples to illustrate the process of the second controller 500 reading the first record R1 in the first recording area 310 to update the second metadata area 510 . The second controller 500 obtains the sequence in which the first data is written into the data storage medium 200 according to the first time stamp T1. Therefore, in this embodiment, the update starts from the data whose first time stamp T1 is “1”, and the data is updated according to the first time stamp T1. The size of T1 is sequentially updated to the data whose first timestamp T1 is "14". The second controller 500 obtains the location information of the first data stored in the data storage medium 200 through the first location L1. As mentioned above, the location information may be the entity of the first data stored in the data storage medium 200 according to different embodiments address, logical address, or a combination of physical and logical addresses. With the location information, the second controller 500 can restore the mapping table written by the first controller 400 to the data storage medium 200 with the first data, and update the second metadata area 510 with the mapping table. In some embodiments, the second controller 500 needs to obtain the data size of the first data through the first capacity C1, and based on the information of the data size of the first data, the second controller 500 can perform the aforementioned restoration of the first controller 400 is an operation of writing the first data into the mapping table of the data storage medium 200 . Because, in some embodiments, the operation process between the physical address and the logical address needs to be matched with the information of the data size to obtain the correct location information.

In addition, when the first controller 400 writes the first data into the data storage medium 200 in a log mode, the first controller 400 will update the first metadata area 410, so the first controller 400 can use the first metadata The area 410 discriminates the data storage order in the data storage medium 200 .

Table 5 is an example of the data storage medium 200 storing data in some embodiments of the present invention. Table 6 is an example of storing records in the first recording area 310 in some embodiments of the present invention. Table 7 is an example of storing records in the second recording area 320 in some embodiments of the present invention. Table 8 is an example of how the first controller 400 and the second controller 500 determine the data storage sequence in the data storage medium 200 according to the records in the recording medium 300 in some embodiments of the present invention. Tables 5 to 8 are as follows:

table 5: Data storage medium 200 Physical block PB1 Physical block PB2 Physical block PB3 Physical block PB4 Physical block PB5 Physical block PB6 Physical block PB7 Physical block PB8 Strip S1 D1 D2 D4 D5 D6 D7 D8 P1 Strip S2 Stripe S3 D3 D5 D6 D7 D8 P3 Stripe S4

Table 6: first recording area 310 first timestamp T1 first position L1 first capacity C1 1 (S1, PB1) d1 2 (S1, PB2) d2 4 (S1, PB3) d4 5 (S1, PB4) d5 8 (S1, PB5) d6 10 (S1, PB6) d7 11 (S1, PB7) d8 . . . . . . . . .

Table 7: The second recording area 320 second time stamp T2 second position L2 The second capacity C2 3 (S3, PB1) d3 6 (S3, PB2) d5 7 (S3, PB3) d6 9 (S3, PB4) d7 12 (S3, PB5) d8 . . . . . . . . .

Table 8: Data storage medium 200 Physical block PB1 Physical block PB2 Physical block PB3 Physical block PB4 Physical block PB5 Physical block PB6 Physical block PB7 Physical block PB8 Strip S1 D1 D2 D4 D5 - old D6-new D7-new D8 - old P1 Strip S2 Stripe S3 D3 D5-new D6 - old D7 - old D8 - new P3 Stripe S4

Referring to FIG. 1 and Tables 5 to 8, in some embodiments, the data storage medium 200 is a fourth-level disk array, the stripes include four stripes S1 to S4, and the physical blocks include eight groups of entities PB1 to PB8 block, and the physical block PB8 is used to store the error correction code. It is assumed that neither the data storage medium 200 nor the recording medium 300 originally stored data or records. Both the first controller 400 and the second controller 500 operate in an active state.

Continuing from the above, the data access system 10 sequentially writes seven first data (respectively D1, D2, D4, D5, D6, D7, and D8 according to the writing order) into the data storage medium 200 through the first controller 400 , and the data access system 10 also sequentially writes five second data (respectively D3, D5, D6, D7, and D8 according to the writing order) into the data storage medium 200 through the second controller 500 . Seven of the first data are written into the strip S1, and five of the second data are written into the strip S3. Therefore, the first data and the second data stored in the data storage medium 200 are shown in Table 5, wherein the error correction code P1 and the error correction code P3 are the error correction codes corresponding to the stripe S1 and the stripe S3, respectively.

The first record R1 in the first recording area 310 is shown in Table 6, and the second record R2 in the second recording area 320 is shown in Table 7. The corresponding records are similar to the descriptions in the foregoing Table 2, and are not described here. Repeat. It should be noted that the first time stamp T1 in Table 6 (1, 2, 4, 5, 8, 10, 11 in sequence) and the second time stamp T2 in Table 7 (3, 6, 7, 9, 12), since the first time stamp T1 and the second time stamp T2 are both generated when the connection interface 100 with multi-path input and output is in the same transmission stage, the first time stamp T1 and the second time stamp T2 The values represented by the two time stamps T2 can be compared with each other, and the larger value represents the newer data.

Therefore, when the first controller 400 reads the second recording area 320 , the first controller 400 can determine the data storage sequence in the data storage medium 200 according to the records in the first metadata area 410 and the second recording area 320 , as shown in Table 8. Similarly, when the second controller 500 reads the first recording area 310 , the second controller 500 can determine the data storage sequence in the data storage medium 200 according to the records in the second metadata area 510 and the first recording area 310 , as shown in Table 8.

To sum up, the data access system of some embodiments of the present application includes a first controller and a second controller. Through the recording in the recording area, another controller can update the metadata area accordingly, so that the two controllers can synchronize their knowledge of the data in the data storage medium. Therefore, the data access system does not need to be limited by the transmission bandwidth between the dual controllers, nor does it need to rely on the mutual exclusion lock mechanism between the dual controllers, so that the data access system can achieve high-speed access controlled by the dual controllers. Function.

10: Data Access System 100: Connection interface 200: Data storage medium 300: Recording Media 310: First recording area 320: Second recording area 400: First Controller 410: First metadata area 500: Second Controller 510: Second metadata area R1: First record R2: Second record T1: first timestamp T2: Second timestamp L1: first position L2: Second position C1: first capacity C2: Second capacity PB1-PB4: Physical Blocks S1-S4: Strip 211-214: Pages 221-224: Pages 231-234: Pages 241-244: Pages 20: Client 21: Connection interface

[FIG. 1] is a schematic diagram of a data access system according to some embodiments of the present invention. 2 is a schematic diagram of a data storage medium according to some embodiments of the present invention. [FIG. 3] is a schematic diagram of a recording medium according to some embodiments of the present invention.

10: Data Access System 100: Connection interface 200: Data storage medium 300: Recording Media 310: First recording area 320: Second recording area 400: First Controller 410: First metadata area 500: Second Controller 510: Second metadata area 20: Client 21: Connection interface

Claims (9)

A data access system includes: a data storage medium; a recording medium including a first recording area and a second recording area; a first controller connected to the data storage medium and the recording medium, and communicating with the data storage medium and the recording medium Corresponding to the first recording area, the first controller includes a first metadata area; and a second controller, connected to the data storage medium and the recording medium, and corresponding to the second recording area, the first The two controllers include a second metadata area; wherein the first controller receives a first data, and updates the first metadata area by writing the first data into the data storage medium in a log mode , and correspondingly generate a first record in the first record area, the second controller updates the second metadata area according to the first record in the first record area; wherein, the first controller checks the When the first record area is full, trigger the second controller to update the second metadata area according to the first record in the first record area, and confirm that the second controller clears the first record area after the update is completed . The data access system of claim 1, wherein when the second controller checks that the occupied capacity of the first record in the first recording area satisfies a threshold, The record updates the second metadata area. The data access system of claim 1, wherein the second controller receives a second data to update the second metadata area by writing the second data into the data storage medium in the log mode , and correspondingly generate a second record in the second record area, and the first controller updates the first metadata area according to the second record in the second record area. The data access system as claimed in claim 3, wherein when the first controller checks that the occupied capacity of the second record in the second recording area satisfies a threshold, according to the second record in the second recording area The record updates the first metadata area. The data access system of claim 3, wherein when the second controller checks that the second recording area is full, triggers the first controller to update the first recording area according to the second record in the second recording area The metadata area is confirmed, and the second recording area is cleared after confirming that the first controller is updated. The data access system of claim 3, further comprising: a multi-path input and output, connected to the first controller and the second controller, the first controller receives the first controller via the multi-path input and output data, the second controller receives the second data via the multi-path input and output. The data access system of claim 6, wherein the first record comprises: a first time stamp of receiving the first data via the multipath input and output, a first time stamp of the first data stored in the data storage medium a location, and a first volume of the first data; the second record includes: a second timestamp of the second data received via the multipath input and output, a second data stored in the data storage medium the second location, and a second volume of the second data. The data access system of claim 1, further comprising: a multi-path input and output connected to the first controller, and the first controller receives the first data via the multi-path input and output. The data access system of claim 8, wherein the first record comprises: a first time stamp of the first data received via the multipath input and output, a first time stamp of the first data stored in the data storage medium a location, and a first volume of the first data.

Images