Patent Application
20140188912
This application is based upon and claims the benefit of priority of the prior Japanese Patent Application No. 2012-289113, filed on Dec. 28, 2012, the entire contents of which are incorporated herein by reference.
The embodiment discussed herein is related to a storage apparatus, a control method, and a computer product.
Conventionally, there has been a bit string data structure called a Bloom Filter. The Bloom filter is used when efficient determination of whether given data is included in an existing data set is performed. According to a related technology, if a first bit string corresponding to a first section to be identified by a search value and range information does not satisfy a generation condition, then input data is registered in the first bit string and, if the first bit string satisfies the generation condition, then the range information and a second bit string of a second section are generated. There is a technology in which in a system of database peers hierarchically connected, a given device has for each lower-layer device, a bit string that indicates the presence of a file set managed by the device of a layer lower than that of the given device; the given device further has a bit string indicating the presence of a file set managed thereby (see, for example, Japanese Laid-Open Patent Publication Nos. 2010-266952 and 2008-102795).
According to the conventional technologies, however, when redundancy determination of determining whether data of the same contents as given data is in an existing data set is performed using a Bloom filter, the load of the redundancy determination increases as the number of Bloom filters increases.
According to an aspect of an embodiment, a storage apparatus includes a storing unit that stores a Bloom filter in which data characteristic values are registered, the data characteristic values extracting properties of data that are stored in areas into which a storage area is divided; and a processor that is configured to judge whether a first data characteristic value extracting a property of a first data that is to be written into the storage area is registered in the Bloom filter, and write the first data into the storage area, when the first data characteristic value is not registered in the Bloom filter.
The object and advantages of the invention will be realized and attained by means of the elements and combinations particularly pointed out in the claims.
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory and are not restrictive of the invention.
Embodiments of a storage apparatus, a control method, and a control program will be described in detail with reference to the accompanying drawings.
For example, the storage system 100 is accessed by a Web server to store Web contents that are to be provided by the Web server to the user. For example, the storage system 100 stores a file to be used by the user.
To suppress the amount of storage of the volume 102, the storage apparatus 101 executes a deduplication technique. The storage apparatus 101, which executes the deduplication technique, performs the following processing with respect to a writing process and a reading process.
With respect to the writing process, the storage apparatus 101 divides the data to be written into blocks. The storage apparatus 101 then calculates for extracted properties of the blocks, data characteristic values that express the characteristics as a specific value. The data characteristic value is, for example, a secure hash value that makes alteration of a block is difficult without alteration of the data characteristic value. As for an algorithm to calculate the secure hash value, there are a message-digest 5 (MD5), a secure hash algorithm (SHA)-1, an SHA-256, etc. Hereinafter, description will be made on the assumption that the data characteristic value is the secure hash value.
The storage apparatus 101 compares the calculated secure hash value and a secure hash value of the block already stored in the volume 102 to judge whether the data is existing data or new data. If the data is existing data, the storage apparatus 101 performs deduplication by not writing the block into the volume 102. If the data is new data, the storage apparatus 101 assigns a physical address of a writing destination within the volume 102 and writes the block thereto. The storage apparatus 101 correlates and adds the calculated secure hash value and the assigned physical address to an index that is for searching for the physical address from the secure hash value. The storage apparatus 101 correlates and stores a logical address and the secure hash value to a correspondence table.
With respect to the reading process, the storage apparatus 101 selects, from the correspondence table, which stores logical addresses correlated with secure hash values, the secure hash value of the block to be read. The storage apparatus 101 then uses the secure hash value of the block to be read and identifies the physical address by referring to the index to search for the physical address from the secure hash value. The storage apparatus 101 reads in the contents of the block to be read, from the identified physical address.
In the reading process and the writing process described above, the index for searching for the physical address from the secure hash value becomes enormous. Because of the use of the secure hash value, records within the index have little locality and for example, even if some records are stored in the memory, shuffling of the records is caused to happen frequently and processing performance drops. Accordingly, use of a Bloom filter can reduce the amount of data of the index. When the bit of the Bloom filter is ON, it indicates a positive or a false positive and when the bit is OFF, it indicates a negative. Configuration may be such that a bit value of 1 will be treated as ON and a bit value of 0 as OFF and conversely, configuration may be such that a bit value of 0 will be treated as ON and a bit value of 1 as OFF. In this embodiment, a bit value of 1 is treated as ON and a bit value of 0 as OFF.
There is a technology of narrowing down the range of a search by preparing plural Bloom filters and determining in which Bloom filter data is hit. The technology of using plural Bloom filters will be described later in
Therefore, the storage apparatus 101 prepares plural small Boom filters and stores in the memory, some of the Bloom filters in which data is predicted to be hit. This makes it possible to suppress the amount of processing required for the deduplication while performing the deduplication to a given extent. The storage capacity of the memory can also be reduced.
In
To select the area predicted to be hit, it is preferable to follow the cache algorithm. For example, the storage apparatus 101 selects an area recently accessed or an area frequently accessed. It is highly likely that a Bloom filter in which the secure hash values of blocks written to each area are registered will include the secure hash value of a block divided from the same file. Therefore, a Bloom filter in which the secure hash values of the blocks written to the areas are registered has high locality. Because of the high locality, it is highly likely that the Bloom filter, in which the secure hash value of given data to be written has been hit, will yield a hit of the secure hash value of subsequent data to be written. The storage apparatus 101 may divide the storage area of the index that is for searching for the physical address from the secure hash value. In the example of
The storage apparatus 101, upon acceptance of a block to be written 106, calculates a secure hash value 107 of the block to be written 106. In the example of
If it is judged that the data characteristic value of the same contents as the secure hash value 107 is not registered in the Bloom filter 105, the storage apparatus 101 writes the block to be written 106 into the area 104-2. Thus, by allowing a given amount of duplicate data, the storage apparatus 101 can suppress the amount of processing required for the deduplication while performing a given degree of deduplication. Details will be described of the storage apparatus 101 with reference to
The user terminals 201#1 to 201#n are clients that use the storage system 100. For example, the user terminals 201#1 to 201#n are typically personal computers (PCs) and use application software such as a Web browser to connect to the storage apparatus 101 and use the storage system 100. The application software is hereinafter referred to as “application”.
The CPU 301 is a computation processing apparatus that governs overall control of the storage apparatus 101. The ROM 302 is non-volatile memory that stores programs such a boot program. The RAM 303 is volatile memory used as a work area of the CPU 301.
The disk drive 304, under the control of the CPU 301, controls the reading and writing of data with respect to the disk 305. For example, a magnetic disk drive, a solid state drive, and the like may be adopted as the disk drive 304. The disk 305 is non-volatile memory that stores data written thereto under the control of the disk drive 304. For example, when the disk drive 304 is a magnetic disk drive, the disk 305 may be a magnetic disk. Further, when the disk drive 304 is a solid state drive, the disk 305 may be semiconductor memory.
The communication interface 306 is a control apparatus that administers an internal interface with the network 202 and controls the input and output of data with respect to other apparatuses. For example, a modem or a LAN adaptor may be employed as the communication apparatus 306. Further, the storage apparatus 101 may have an optical disk drive, an optical disk, a keyboard, and a mouse.
With reference to
A fifth-tier Bloom filter includes a BF5-1 and a BF5-2 as subordinates of the BF4-1, a BF5-3 and a BF5-4 as subordinates of the BF4-2, a BF5-5 and a BF5-6 as subordinates of the BF4-3, and a BF5-7 and a BF5-8 as subordinates of the BF4-4. The fifth-tier Bloom filter further includes a BF5-9 and a BF5-10 as subordinates of the BF4-5 and a BF5-11 and a BF5-12 as subordinates of the BF4-6. The fifth-tier Bloom filter further includes a BF5-13 and a BF5-14 as subordinates of the BF4-7 and a BF5-15 and a BF5-16 as subordinates of the BF4-8.
If the hash value of data searched-for is not hit in a test of the first-tier Bloom filter BF1-1, then the storage apparatus 101 judges that the data searched-for is not included.
If the data searched-for is hit in the test of the Bloom filter BF1-1, the storage apparatus 101 judges if the data searched-for is hit in the second-tier Bloom filter BF2-1. If the BF2-1 does not yield a hit, the storage apparatus 101 searches to see if the data searched-for is included in the Bloom filter BF2-2. Thus, if a Bloom filter yields a hit, the storage apparatus 101 performs the test of subordinate Bloom filters thereof and narrows down the range of search, thereby reaching the target data.
The Bloom filter, which has the false positive, can cause an erroneous detection. In the case of erroneous detection, the storage apparatus 101 is only required to go back to the superior Bloom filter and test the untested Bloom filter. Suppose, for example, that the storage apparatus 101 gets a hit in the test of the BF4-1 but does not get a hit in the test of the BF5-1 or the BF5-2. In this case, the hit in the test of the BF4-1 is the false positive and the storage apparatus 101 goes back to the superior Bloom filter and performs the test of the BF4-2, which is next.
The storage apparatus 101 can reduce the number of tiers by controlling the number of divisions.
In the example of
Functions of the storage apparatus 101 will be described.
The storage apparatus 101 can access a storing unit 620. The storing unit 620 is storage devices such as the RAM 303 and the disk 305. The storing unit 620 includes a block map table 621, a writing object MBF cache 622, an MBF cache table 623, an MBF table 624, and a hash log table 625. The writing object MBF cache 622 and the MBF cache table 623 reside in the memory serving as a main storage unit such as the RAM 303 and a register, a cache memory, etc. of the CPU 301. The block map table 621, the MBF table 624, and the hash log table 625 reside in a disk serving as an auxiliary storage unit such as the disk 305.
The block map table 621 correlates and stores the secure hash value of each block to be written and the logical address of each data to be written, corresponding to each data to be written among a group of data to be written. Details of the block map table 621 will be described in
The writing object MBF cache 622 and the MBF cache table 623 store the Bloom filter that has the secure hash value of the block stored in at least any one area selected from among areas into which the storage area of the volume 102 is divided. The MBF cache table 623 may include the writing object MBF cache 622 within itself. The MBF cache table 623 may store the Bloom filter that has the secure hash value of the block stored in at least any one area selected from among areas into which the storage area of the hash log table 625 is divided. In this embodiment, description will be made using the example of dividing the storage area of the hash log table 625.
The MBF table 624 stores the Bloom filter that has for each area, the secure hash value of the block stored in the area. Details of the writing object MBF cache 622 to the MBF table 624 will be described later in
The hash log table 625 corresponds to the index that is for searching for the physical address from the secure hash value described in
The writing judging unit 601 judges whether a secure hash value of the same contents as that of a first secure hash value of the data to be written to the storage area of the hash log table 625 is not registered in the Bloom filter of the MBF cache table 623. The Bloom filter in the MBF cache table 623 may be one Bloom filter or may be plural Bloom filters. Results of the judgment are stored in the register and the cache memory of the CPU 301, the RAM 303, etc.
If it is judged by the writing judging unit 601 that a secure hash value of the same contents is not registered, the determining unit 602 determines whether to register the first secure hash value in the Bloom filter of the writing object MBF cache 622, based on the following condition. The “following condition” indicates the number of secure hash values already registered in the writing object MBF cache 622.
If another Bloom filter is acquired by the acquiring unit 603, the determining unit may determine whether to register the first secure hash value in the other Bloom filter, based on the number of the secure hash values already registered in the other Bloom filter. Results of the determination are stored in the register and the cache memory of the CPU 301, the RAM 303, etc.
If the determining unit 602 determines not to register the first secure hash value in the Bloom filter, the acquiring unit 603 acquires from the storing unit 620, another Bloom filter different from the Bloom filter of the MBF cache table 623. The other Bloom filter may be a Bloom filter newly prepared in the storing unit 620 or may be a Bloom filter that has not reached the registration upper limit number among the Bloom filters of the MBF table 624. Results of the acquisition are stored in the register and the cache memory of the CPU 301, the RAM 303, etc.
If it is judged by the writing judging unit 601 that a secure hash value of the same contents is registered, the writing detecting unit 604 detects the data having the secure hash value of the same contents from at least any one area. Results of the detection are stored in the register and the cache memory of the CPU 301, the RAM 303, etc.
If it is judged by the writing judging unit 601 that a secure hash value of the same contents is not registered, the writing unit 605 writes the data to be written to the storage area of the volume 102. If, because of the division of the storage area of the hash log table 625, it is judged by the writing judging unit 601 that a secure hash value of the same contents is not registered, the writing unit 605 writes to the hash log table 625, the physical address at which the data that is to be written is stored.
If the determining unit 602 determines to register the first secure hash value in the Bloom filter, the writing unit 605 may write to at least any one area, the data that is to be written.
If the determining unit 602 determines to register the first secure hash value in another Bloom filter, the writing unit 605 may write to an area that stores data having a secure hash value registered in the other Bloom filter, the data that is to be written. If data having a secure hash value of the same contents is detected by the writing detecting unit 604, the writing unit 605 need not write to the storage area of the volume 102, the data that is to be written. If data having a secure hash value of the same contents is not detected by the writing detecting unit 604, the writing unit 605 may write to the storage area of the volume 102, the data that is to be written.
If the determining unit 602 determines to register the first secure hash value in the Bloom filter, the registering unit 606 registers the first secure hash value in the Bloom filter.
The updating unit 607 updates the contents of the writing object MBF cache 622, based on the other Bloom filter acquired by the acquiring unit 603. For example, the updating unit 607 saves the data of the writing object MBF cache 622 to the MBF cache table 623 and overwrites the writing object MBF cache 622 with the other Bloom filter. At the time of saving the writing object MBF cache 622 to the MBF cache table 623, if there is an empty area in the MBF cache table 623, the updating unit 607 writes the data of the writing object MBF cache 622 in the empty area. If there is no empty area in the MBF cache table 623, the updating unit 607 overwrites an old record with the other Bloom filter.
The updating unit 607 may update the contents of the writing object MBF cache 622, based on the Bloom filter that has the data having a secure hash value of the same contents as that of a second secure hash value identified by the identifying unit 613.
The selecting unit 611 selects from the block map table 621, the second secure hash value correlated with the logical address of the data that is to be read. Results of the selection are stored in the register and the cache memory of the CPU 301, the RAM 303, etc.
The reading judging unit 612 judges whether a secure hash value of the same contents as that of the second secure hash value selected by the selecting unit 611 is not registered in the Bloom filter of the MBF cache table 623. Results of the judgment are stored in the register and the cache memory of the CPU 301, the RAM 303, etc.
In the following case, the identifying unit 613 identifies the Bloom filter that has data having a secure hash value of the same contents as that of the second secure hash value, from among the Bloom filters of the MBF table 624. The “following case” indicates a case in which it is judged by the reading judging unit 612 that a secure hash value of the same contents as that of the second secure hash value is not registered. Results of the identification are stored in the register and the cache memory of the CPU 301, the RAM 303, etc.
In the following case, the reading detecting unit 614 detects from at least any one area, data having a secure hash value of the same contents as that of the second secure hash value. The “following case” indicates a case in which it is judged by the reading judging unit 612 that a secure hash value of the same contents as that of the second secure hash value is registered.
The reading detecting unit 614 may detect a data having a secure hash value of the same contents as that of the second secure hash value, from the area in which the data is stored that has the secure hash value registered in the Bloom filter identified by the identifying unit 613. Results of the detection are stored in the register and the cache memory of the CPU 301, the RAM 303, etc.
If data having a secure hash value of the same contents as that of the second secure hash value is detected by the reading detecting unit 614, the output unit 615 outputs the data having the secure hash value of the same contents as that of the second secure hash value. Output may be to a storage area of the RAM 303, the disk 305, etc., or may be to the application of the user terminal 201 that made a read request.
For example, for the block indicated by record 701-1, the volume ID of the volume storing the block is 1, the logical block address is 0, the MBF-ID of the object block is 0, and the secure hash value of the object block is 0xe251e71 . . . .
The MBF-ID field stores the identification number of the Bloom filter. The MBF index data field stores a bit string as the Bloom filter. One MBF index data field may store plural Bloom filters. For example, in
For example, record 801-1 indicates that the MBF-ID is 3 and that the bit string as the Bloom filter is “zzzzzzzz . . . ”.
The hash log table 625 comes to have an enormous number of records and therefore, to narrow down the range of a search, the contents of the hash log table 625 are divided, for each Bloom filter for which the MBF-ID is hit. The range of search narrowed down by the Bloom filters that yield a hit is hereinafter referred to as “hash log range”. In the example of
For example, it is assumed that the BF2-1 and the BF3-1 to BF3-4 are stored in the MBF index data field of record 803-1 depicted in
An operation of the reading process and the writing process of the storage apparatus 101 will be described with reference to
The storage apparatus 101 accepts from the application, a request to read the block having the volume ID of “1” and the logical block address of “2”. The storage apparatus 101 then detects the record having the volume ID of “1” and the logical block address of “2” in the block map table 621. In the example of
The storage apparatus 101 then searches for a record having “1” as the value of the MBF-ID field in the MBF cache table 623. In the example of
In the example of
The storage apparatus 101 accepts from the application, a request to read the block having the volume ID of “1” and the logical block address of “0”. The storage apparatus 101 then detects a record having the volume ID of “1” and the logical block address value of “0” in the block map table 621. In the example of
The storage apparatus 101 then searches for a record having “0” as the value of the MBF-ID field from the MBF cache table 623. In the example of
As to which BF is hit, in the example of
The storage apparatus 101 accepts a request to write block b1 having the volume ID of “1”, the logical block address of “3”, and the data contents of “0x01234567”. The storage apparatus 101 then calculates the secure hash value of “0x01234567”. In the example of
In the example of
For example, since the same processing as the reading process described in
After confirming that there is the record having the secure hash value “0xe251eb71 . . . ”, the storage apparatus 101, using the contents of the write request, updates the block map table 621. For example, if there is the record having the volume ID field value of “1” and the logical block address of “3 in the block map table 621, the storage apparatus 101 updates the applicable record and if there is no such record, the storage apparatus 101 adds the record. If there is the applicable record, the storage apparatus 101 updates the MBF-ID field of the applicable record with the value “0” of the MBF-ID field of the detected record 802-3 as well as updating the secure hash value with “0xe251eb71 . . . ”. The example of
The storage apparatus 101 accepts a request to write block b2 having the volume ID of “1”, the logical block address of “4”, and the data contents of “0x13572468”. The storage apparatus 101 then calculates the secure hash value of “0x13572468”. In the example of
In the example of
The storage apparatus 101 acquires the value “3” of the MBF-ID field of record 801-1 and hash log range 911-4 specified by the BF in which the registration is made. The storage apparatus 101 adds record 904-1 having the value “0x5541a022 . . . ” of the secure hash value field and the value “5” of the physical block address field, as a record to be included in hash log range 911-4. The storage apparatus 101, using the contents of the write request, updates the block map table 621. In the example of
A flowchart of the operation of the reading process and the writing process described using
If a record could be detected (step S1402: YES), then the storage apparatus 101 acquires the MBF-ID of the detected record (step S1403). The storage apparatus 101, using the acquired MBF-ID as a key, searches for a matching record from the MBF cache table 623 (step S1404). The storage apparatus 101 judges if a record could be detected (step S1405). If a record could not be detected (step S1405: NO), the storage apparatus 101, using the acquired MBF-ID as a key, searches for a matching record in the MBF table 624 (step S1406). The storage apparatus 101 replaces the record of the MBF cache table 623 by the detected record (step S1407).
After the completion of execution of step S1407 or if a record could be detected (step S1405: YES), the storage apparatus 101 acquires the hash log range from the MBF-ID and the MBF index data of the detected record (step S1408). The storage apparatus 101, using the secure hash value found in the block map table 621 as a key, searches for a matching record among the group of records of the hash log table 625 included in the acquired hash log range (step S1409). The storage apparatus 101 outputs the contents of the block to be read from the physical block address of the detected record (step S1410).
If a record could not be detected (step S1402: NO), the storage apparatus 101, judging that the requested block is not a block written by the application, outputs data embedded with 0 (step S1411). After the completion of execution of step S1410 or step S1411, the storage apparatus 101 ends the reading process. With the execution of the reading process, the storage apparatus 101, by using the BF to the storage to which the deduplication technology was applied, can perform high-speed judgment of the storage location of the block and therefore, can perform the reading of the block contents at high speed.
The storage apparatus 101 calculates the secure hash value from the contents of the block of the accepted write request (step S1501). The storage apparatus 101 searches for a record having the calculated secure hash value in the MBF index data, in the MBF cache table 623 (step S1502). The storage apparatus 101 judges if a record could be detected (step S1503). If a record could not be detected (step S1503: NO), the storage apparatus 101 proceeds to the operation at step S1601 depicted in
If a record could be detected (step S1503: YES), the storage apparatus 101 acquires the hash log range from the MBF-ID and the MBF index data of the detected record (step S1504). The storage apparatus 101, using the calculated secure hash value as a key, searches for a matching record among the group of records of the hash log table 625 included in the acquired hash log range (step S1505). The storage apparatus 101 judges if a record could be detected (step S1506). If a record could not be detected (step S1506: NO), storage apparatus 101 goes to the operation at step S1601 depicted in
If the record could be detected (step S1506: YES), then the storage apparatus 101 acquires the MBF-ID of the record detected from the MBF cache table 623 (step S1507). After the completion of the operation at step S1507, the storage apparatus 101 updates the block map table 621 with a set of the volume ID and the logical block address of the accepted write request, the acquired MBF-ID, and the calculated secure hash value (step S1508). After the completion of the operation at step S1608 depicted in
In the case of step S1503: NO or step S1506: NO, the storage apparatus 101 judges if the number of registrations of the writing object MBF cache has reached a predetermined number (step S1601). If the number of registrations of the MBF cache has reached the predetermined number (step S1601: YES), the storage apparatus 101 writes the writing object MBF cache 622 to the MBF cache table 623 (step S1602). With respect the operation at step S1602, old records present in the MBF cache table 623 are directly discarded since the same contents are in the MBF table 624. The storage apparatus 101 creates the MBF index data with all bits OFF (step S1603). The storage apparatus 101 establishes a set of a new MBF-ID and the created MBF index data as a new writing object MBF cache 622 (step S1604). In the operation at step S1603 and step S1604, if there is another BF that has not reached the upper limit number, the storage apparatus 101 may establish the BF as a new writing object MBF cache 622, without executing the operation at step S1603.
After the completion of the operation at step S1604 or if the number of registrations of the MBF cache has not reached the predetermined number (step S1601: NO), the storage apparatus 101 acquires the MBF-ID of the writing object MBF cache 622 (step S1605). The storage apparatus 101 then writes the block contents of the accepted write request to the volume (step S1606). The storage apparatus 101 registers the calculated secure hash value in the MBF index data of the writing object MBF cache 622 (step S1607). The storage apparatus 101 adds the calculated secure hash value and the physical block address of the written block contents as a record to the hash log range specified by the acquired MBF-ID and the registering BF, in the hash log table 625 (step S1608). After the completion of the operation at step S1608, the storage apparatus 101 proceeds to step S1508.
With the execution of the writing process, the storage apparatus 101 can reduce the amount of processing required for the deduplication while performing a given degree of deduplication, using the BF. A comparison of the performance according to this embodiment will be described with reference to
The four trace data are described below. First trace data “iodedup.homes” describes I/O patterns of the home directory. Second trace data “iodedup.mail” describes the I/O patterns of the mail server. Third trace data “srcmap.home1” describes frames and events transmitted to home1. Fourth trace data “srcmap.home2” describes the frames and the events transmitted.
As shown by graph 1701, when there are one billion data, if the memory has one million data that account for one-thousandth thereof, the storage apparatus 101 can detect about 90% of the duplication.
As depicted in
As described above, according to the storage apparatus 101, at the time of the block writing, the MBF cache table 623 residing in the memory is searched rather than the MBF table 624 residing in the disk is searched, and if the block is not registered, the block is written. This makes it possible to reduce the load of the duplication judgment while performing the deduplication of the data to a given extent.
According to the storage apparatus 101, at the time of the block writing, the Bloom filter of the MBF cache table 623 is tested and if the block is not registered, the block is registered in the writing object MBF cache 622. Since the writing object MBF cache 622 registers the secure hash values of the blocks to be written occurring around the same time, the storage apparatus 101 increases the locality of the writing object MBF cache 622. The increase locality reduces the number of accesses of the disk and reduces the processing time required for the deduplication.
According to the storage apparatus 101, when the number of the secure hash values registered in the writing object MBF cache 622 has reached the upper limit, another Bloom filter may be used. This enables the storage apparatus 101 to suppress the false positive judgment rate.
According to the storage apparatus 101, the MBF table may be updated by using another Bloom filter. By this, since the read request or the write request is liable to happen in a short time to a new Bloom filter, the storage apparatus 101 may enhance the cache hit rate.
According to the storage apparatus 101, when the Bloom filter of the MBF cache table 623 is tested and the block to be written is registered, a block of the same value as the secure hash value of the block to be written is searched for and if detected, the block to be written need not be written. This enables the storage apparatus 101 to perform the deduplication and reduce the amount of memory of the volume 102.
According to the storage apparatus 101, when the Bloom filter of the MBF cache table 623 is tested and the block to be written is registered, the block of the same value as the secure hash value of the block to be written is searched for and if not detected, the block to be written may be written. This enables the storage apparatus 101 to hold the contents of the block normally even if the block is erroneously detected by the false positive.
According to the storage apparatus 101, when the Bloom filter of the MBF cache table 623 is tested and the block to be read is registered, the same value as the secure hash value of the block to be read may be searched for from the range of the storage area indicated by the Bloom filter. Since this narrows down the range within which to search, the storage apparatus 101 may output the object data at high speed.
According to the storage apparatus 101, when the Bloom filter of the MBF cache table 623 is tested and the block to be read is not registered, the Bloom filter of the MBF table 624 may be tested. This enables the storage apparatus 101 to output the block to be read normally even if the memory is not hit.
According to the storage apparatus 101, among the Bloom filters of the MBF table 624, the Bloom filter that is hit may be set in the MBF cache table 623. By this, since such secure hash values that will increase the locality are registered in the Bloom filters, the storage apparatus 101 can estimate that a read request to or a write request to is liable to occur in a short time and can enhance the cache hit rate.
When the MBF cache table 623 is not used, if, for example, maximum 10 [TB] area is managed by 4 [KB] block, the storage apparatus not using the MBF cache table 623 manages 25 million blocks. The storage apparatus using the two-tier MBF using 23 [bit] per block secures the memory of 2.5 [MB]×23×2 [bit]=around 14 [GB]. By contrast, in this embodiment, the size of the area and the size of the installed memory become irrelevant and even when, for example, 10 [TB] area is managed by 4 [KB] block, the storage apparatus 101 is operable even with the memory of 1 [GB] or less.
The control method described in the present embodiment may be implemented by executing a prepared program on a computer such as a personal computer and a workstation. The program is stored on a non-transitory, computer-readable recording medium such as a hard disk, a flexible disk, a CD-ROM, an MO, and a DVD, read out from the computer-readable medium, and executed by the computer. The program may be distributed through a network such as the Internet.
According to one aspect of an embodiment, reductions in the load of the redundancy determination using the Bloom filter are enabled.
All examples and conditional language provided herein are intended for pedagogical purposes of aiding the reader in understanding the invention and the concepts contributed by the inventor to further the art, and are not to be construed as limitations to such specifically recited examples and conditions, nor does the organization of such examples in the specification relate to a showing of the superiority and inferiority of the invention. Although one or more embodiments of the present invention have been described in detail, it should be understood that the various changes, substitutions, and alterations could be made hereto without departing from the spirit and scope of the invention.
| Number | Date | Country | Kind |
|---|---|---|---|
| 2012-289113 | Dec 2012 | JP | national |
