1. Technical Field
This disclosure relates to disk drives and data storage systems for computer systems. More particularly, the disclosure relates to systems and methods for reducing the size of translation tables used by a disk drive or other data storage system.
2. Description of the Related Art
Disk drives typically comprise a disk and a head connected to a distal end of an actuator arm which is rotated about a pivot by a voice coil motor (VCM) to position the head radially over the disk. The disk comprises a plurality of radially spaced, concentric tracks for recording user data sectors and embedded servo sectors. The embedded servo sectors comprise head positioning information (e.g., a track address) which is read by the head and processed by a servo controller to control the velocity of the actuator arm as it seeks from track to track.
Disk drive capacity has grown by nearly six orders of magnitude since the introduction of magnetic disk storage in 1956. Current magnetic disk drives that use perpendicular recording are capable of storing as much as 400 Gbit/in2. However, this is rapidly approaching the density limit of about 1 Tbit/in2 (1000 Gbit/in2) due to the superparamagnetic effect for perpendicular magnetic recording. Accordingly, new approaches for improving storage density are needed.
Systems and methods that embody the various features of the invention will now be described with reference to the following drawings, in which:
While certain embodiments are described, these embodiments are presented by way of example only, and are not intended to limit the scope of protection. Indeed, the novel methods and systems described herein may be embodied in a variety of other forms. Furthermore, various omissions, substitutions and changes in the form of the methods and systems described herein may be made without departing from the scope of protection.
Overview
Shingled writing can improve the density of magnetic disk storage. Because writing data to a magnetic medium generally requires a stronger magnetic field than reading data from the magnetic medium, adjacent data tracks can be partially overlapped or “shingled” during writing. This can be achieved by using a wider and stronger magnetic field when writing data as compared to the magnetic field used when reading data. As a result, shingled writing allows for narrower tracks, which can increase the data density by a factor of about 2.3 or higher. Disk drives that utilize shingled writing are referred to as “shingled disk drives” or “shingled hard drives” throughout this disclosure.
While shingled disk drives permit random access data reading, writing is performed sequentially because writing to a single track can affect (e.g., partially or fully overwrite) data in one or more adjacent tracks. For some shingled disks, the number of affected neighboring tracks can be between 4 and 8. Accordingly, shingled disks drives provide random read access and sequential write access.
To prevent inter-track interference, shingled disk drives can be configured to write data in tracks sequentially in one direction (e.g., from the outer diameter toward the inner diameter of a disk platter) so that a previously written track is overwritten only by the next adjacent track. A shingled disk drive can implement a log-structured file system (LFS), in which write operations progress in the same radial direction so that the data tracks can overlap. New data can be written to the head of a circular buffer (e.g., current data track). During garbage collection, valid data can be relocated from the tail of the buffer (e.g., preceding data tracks) to the head of the buffer in order to free space for new data to be written.
Shingled disk drives can further utilize one or more translation tables that map logical addresses used by a host system to access a disk drive to physical locations or addresses in the disk drive. For example, the host can access the disk drive (e.g., to store and/or retrieve data) as a linear array of logical block addresses (LBAs), and the disk drive can utilize a translation table to map the logical block addresses to physical locations on a magnetic disk where host data is stored. Translation tables can be used by the disk drive to locate host data stored in the drive. In addition, the disk drive can be configured to store metadata information used by the disk drive to keep track of locations where host data is stored. For example, metadata for a particular track can include the mapping of logical addresses for which host data is stored on the track to physical addresses where host data is stored. Typically, the host system does not use metadata information. Instead, metadata information is used by the disk drive to locate host data stored in the disk drive.
In some embodiments of the present invention, a translation table of reduced size is maintained by a disk drive system. The disk drive can be configured to utilize an additional address mapping layer between the logical addresses used by the host system and physical locations in the disk drive. This additional mapping layer can relate the logical address used by the host system to another set of logical addresses utilized by the disk drive, which in turn correspond to physical locations in the disk drive. Logical addresses utilized by the disk drive exclude or skip some or all physical locations that are configured to store metadata. Accordingly, entries in the translation table can comprise logical addresses used by the host system along with corresponding logical addresses utilized by the disk drive. This approach permits the translation table to be smaller and, therefore, results in a more efficient storage and retrieval of data.
In some embodiments of the present invention, the disk drive can be configured to further reduce the size of the translation table by utilizing one or more suitable data encoding or compression schemes, such as run length compression or encoding, Huffman coding, etc. For example, the host system can store consecutive or sequential data values starting at a given logical address in the disk drive. The disk drive can be configured to store the sequential data values along with metadata, for example, in the magnetic disk storage beginning with a given physical location. According to some embodiments, the translation table can comprise a single entry that includes the given logical address used by the host system as the starting address, a logical address utilized by the disk drive corresponding to the given physical location where a first data value of the sequence is stored, and a number of data values in the sequence. Using this additional mapping layer that excludes some or all physical locations configured to store metadata allows the disk drive to reduce the size of the translation table by taking advantage of run length encoding when a consecutive run of host data is stored in the disk drive.
In some embodiments of the present invention, the disk drive can be configured to retrieve data requested by the host by using the translation table. The disk drive can determine a logical address utilized by the disk drive that corresponds to the logical address specified by host in connection with a read data operation. The disk drive can convert the logical address utilized by the disk drive to a corresponding physical location in accordance with a predetermined conversion process. For example, when metadata information is stored in predetermined locations in the disk drive (e.g., in starting and middle locations of each track), the conversion process can be a mapping, such as a simple transformation, of logical addresses utilized by the disk drive to physical addresses.
System Overview
In one embodiment, the disk 102 comprises a plurality of servo sectors 240-24N that define the plurality of data tracks 10. The controller 14 processes the read signal to demodulate the servo sectors 240-24N into a position error signal (PES). The PES is filtered with a suitable compensation filter to generate a control signal 26 applied to a voice coil motor (VCM) 110 which rotates the actuator arm 4 about a pivot in order to position the head 104 radially over the disk 102 in a direction that reduces the PES. The servo sectors 240-24N may comprise any suitable position information, and in one embodiment, as is shown in
In one embodiment, as is illustrated in
The controller 230 can be configured to receive data and/or storage access commands from a storage interface module 212 (e.g., a device driver) of a host system 210. Storage access commands communicated by the storage interface 212 can include write and read commands issued by the host system 210. Read and write commands can specify a logical address (e.g., LBA) in the disk drive 220. The controller 230 can execute the received commands in the non-volatile memory module 250 and/or in the magnetic storage module 260.
Disk drive 220 can store data communicated by the host system 210. In other words, the disk drive 220 can act as memory storage for the host system 210. To facilitate this function, the controller 230 can implement a logical interface. The logical interface can present to the host system 210 disk drive's memory as a set of logical addresses (e.g., contiguous address) where host data can be stored. Internally, the controller 230 can map logical addresses to various physical locations or addresses in the magnetic media 264 and/or the non-volatile memory module 250. In some embodiments, the controller 230 can map logical addresses used by the host system to logical addresses utilized by the disk drive (e.g., logical addresses exclude or skip physical locations configured to store metadata information), which in turn correspond to physical locations in the disk drive.
In the embodiment illustrated in
Row C of
In the embodiment illustrated in
In one embodiment, as is illustrated by the table stored in location 404, metadata information includes a mapping between logical addresses for which host data is stored on the half track and physical locations in the half track. Metadata information can be generated and stored in the disk drive (e.g., in the magnetic media 264) by the controller 230 when a write command received from the host is processed (e.g., host data is written to the disk drive). As is illustrated, metadata information includes a mapping for the previous half track. In other words, the mapping stored in location 404 depicts the data layout of the first half of track 0. In particular, the entry in row BB indicates that a sequence of three host data values beginning with LBA 100 (i.e., values corresponding to LBA 100-102) is stored in physical locations starting at ABA 1. Similarly, the entry in row CC indicates that a sequence of two host data values beginning with the next LBA in the sequence (LBA 103) is stored in physical locations starting at ABA 5. As is explained above, no host data is stored in the defective location 406 corresponding to ABA 4. Those of ordinary skill in the art will appreciate that metadata information can include a mapping for the current half track, current track, plurality of current or previous half tracks, tracks, etc. In addition, those of ordinary skill in the art will appreciate other suitable mapping techniques or combinations of such techniques can be used to generate metadata or write log information.
Row C of
Reduced Size Translation Table
With reference to
With reference to
As is illustrated, physical locations configured to store metadata fragment or break-up entries in the translation table 400B. In other words, because metadata is interspersed with host data stored on the disk drive, sequential runs of host data are broken up by physical locations that store metadata. This fragmentation is reflected in table 400B. In addition, entries in table 400B are broken-up by the defective location 406 corresponding to ABA 4.
As is illustrated in
In one embodiment, the translation table (e.g., table 400B, 400C, etc.) can be stored or cached in volatile memory, such as cache buffer 262 (which can be implemented in DRAM in one embodiment), which can provide faster access to the table as compared to the table being stored only in the magnetic media 264. In another embodiment, the translation table can be stored in non-volatile memory 250. The disk drive can be further configured to store a working copy of the translation table in volatile memory, such as cache buffer 262 or non-volatile memory 250, while storing a second (e.g., backup) copy of the translation table in the magnetic media 264. When multiple copies of the translation table are used, the disk drive can be configured to periodically save the working copy of the translation table. For example, the controller 230 can be configured to save the working copy in a reserved area of the magnetic media 264.
In block 506, the process 500 determines a logical address utilized by the disk drive (e.g., ShABA address) that corresponds to the starting LBA address associated with the read data command. In one embodiment, the process 500 can search the translation table 400C (e.g., perform a look-up operation) and determine the logical address corresponding to the starting LBA address. The process transitions to block 508 where it determines the number of consecutive host data values that follow the logical address corresponding to the starting LBA address (including the data value corresponding to the logical address).
In block 510, the process determines whether the count of consecutive data values to be retrieved is smaller than or equal to than the number of consecutive host data values that follow the logical address corresponding to the starting LBA address. If the count of consecutive data values is determined to be smaller or equal than the number of consecutive host data values, the process transitions to block 512. Otherwise, the process transitions to block 514.
In block 512, the process 500 retrieves data values consecutively stored beginning with the logical address corresponding to the starting LBA address. The number of consecutive data values retrieved corresponds to the count specified in block 502. In one embodiment, the process 500 can convert the logical address (e.g., ShABA address) corresponding to the starting LBA address to a physical address (e.g., ABA address) by using a predetermined conversion process. For example, as is explained above, logical addresses utilized by the disk drive can be mapped to corresponding physical addresses in the disk drive 220. Further, when metadata information is stored according to a known or predetermined pattern (e.g., in starting and middle locations of each track), the conversion process can be a simple transformation of logical addresses utilized by the disk drive to physical addresses. For mapping illustrated in
Physical address=Logical address utilized by disk drive+n (1)
where n corresponds to a half track number (selected from the set [1, N]) for the logical address utilized by the disk drive. Starting at the physical address corresponding to the logical address utilized by the disk drive, the process 500 retrieves a number of consecutively stored data values corresponding to the count specified in block 502. When the process completes the retrieval of host data, it transitions to 520 where data is communicated (e.g., returned) to the host system.
If the process 500 determines in block 510 that the count specified in block 502 is greater than the number of consecutive host data values that follow the logical address corresponding to the starting LBA address, the process transitions to block 514. In block 514, the process 500 converts the logical address corresponding to the starting LBA address to a physical address and retrieves host data stored in the disk drive beginning with that physical address. The process retrieves a number of consecutively stored data values corresponding to the number determined in block 508, and updates (e.g., increments) the counter of retrieved data values. The process transitions to block 516 where it determines whether the number of retrieved data values (as tracked by the counter) equals or exceeds the count specified in block 502. If the process determines that all data values requested by the host system have been retrieved, it transitions to block 520 where data is communicated (e.g., returned) to the host system.
If in block 516 the process 500 determines that it needs to retrieve more host data values stored in the disk drive, the process transitions to block 518. In block 518, the process determines the next logical address utilized by the disk drive. In one embodiment, the process 500 can search the translation table 400C and determine such next logical address. For example, the process can determine the next logical address as follows:
starting LBA address specified in block 502+number of consecutively stored data values determined in block 508.
Alternatively, the process 500 can determine the next logical address by looking-up the next entry in the table 400C. The process also determines the number of consecutive host data values that follow the next logical address (including the data value corresponding to the next logical address). The process transitions to block 514 where data is retrieved as explained above.
By using an additional address mapping layer between the logical addresses used by the host system and physical locations in the disk drive, a translation table of reduced size can maintained by the disk drive. In addition, the use of data encoding or compression schemes, such as run length encoding, can further reduce the size of the translation table. Maintaining a reduced-size translation table can result in more efficient data storage and retrieval. In addition, when a working copy of the translation table is stored in volatile memory, such as DRAM, volatile memory of a smaller size can be used, which results in cost savings. Moreover, by excluding or skipping physical locations configured to store metadata information from the additional mapping layer, metadata can be better concealed from the host system. For example, the host system may need to communicate a special (e.g., non-standard command) to access physical locations configured to store metadata information. This can provide better security for data stored in the disk drive.
Those skilled in the art will appreciate that in some embodiments, other techniques to reduce the size of translation tables can be implemented. In addition, the actual steps taken in the disclosed processes, such as the process shown in
While certain embodiments have been described, these embodiments have been presented by way of example only, and are not intended to limit the scope of the protection. Indeed, the novel methods and systems described herein may be embodied in a variety of other forms. Furthermore, various omissions, substitutions and changes in the form of the methods and systems described herein may be made without departing from the spirit of the protection. The accompanying claims and their equivalents are intended to cover such forms or modifications as would fall within the scope and spirit of the protection. For example, the systems and method disclosed herein can be applied to hybrid hard drives, solid state drives, and the like. In addition, other forms of storage (e.g., DRAM or SRAM, battery backed-up volatile DRAM or SRAM devices, EPROM, EEPROM memory, etc.) may additionally or alternatively be used. As another example, the various components illustrated in the figures may be implemented as software and/or firmware on a processor, ASIC/FPGA, or dedicated hardware. Also, the features and attributes of the specific embodiments disclosed above may be combined in different ways to form additional embodiments, all of which fall within the scope of the present disclosure. Although the present disclosure provides certain preferred embodiments and applications, other embodiments that are apparent to those of ordinary skill in the art, including embodiments which do not provide all of the features and advantages set forth herein, are also within the scope of this disclosure. Accordingly, the scope of the present disclosure is intended to be defined only by reference to the appended claims.
Number | Name | Date | Kind |
---|---|---|---|
4769770 | Miyadera et al. | Sep 1988 | A |
4992936 | Katada et al. | Feb 1991 | A |
5121480 | Bonke et al. | Jun 1992 | A |
5293282 | Squires et al. | Mar 1994 | A |
5613066 | Matsushima et al. | Mar 1997 | A |
5983309 | Atsatt et al. | Nov 1999 | A |
6092231 | Sze | Jul 2000 | A |
6105104 | Guttmann et al. | Aug 2000 | A |
6182250 | Ng et al. | Jan 2001 | B1 |
6182550 | Brewington et al. | Feb 2001 | B1 |
6202121 | Walsh et al. | Mar 2001 | B1 |
6240501 | Hagersten | May 2001 | B1 |
6324604 | Don et al. | Nov 2001 | B1 |
6339811 | Gaertner et al. | Jan 2002 | B1 |
6411454 | Monroe, III | Jun 2002 | B1 |
6556365 | Satoh | Apr 2003 | B2 |
6574774 | Vasiliev | Jun 2003 | B1 |
6636049 | Lim et al. | Oct 2003 | B1 |
6690538 | Saito et al. | Feb 2004 | B1 |
6728054 | Chng et al. | Apr 2004 | B2 |
6735032 | Dunn et al. | May 2004 | B2 |
6772274 | Estakhri | Aug 2004 | B1 |
6829688 | Grubbs et al. | Dec 2004 | B2 |
6886068 | Tomita | Apr 2005 | B2 |
6895468 | Rege et al. | May 2005 | B2 |
6901479 | Tomita | May 2005 | B2 |
6920455 | Weschler | Jul 2005 | B1 |
6956710 | Yun et al. | Oct 2005 | B2 |
6967810 | Kasiraj et al. | Nov 2005 | B2 |
6980386 | Wach et al. | Dec 2005 | B2 |
6992852 | Ying et al. | Jan 2006 | B1 |
7012771 | Asgari et al. | Mar 2006 | B1 |
7035961 | Edgar et al. | Apr 2006 | B2 |
7046471 | Meyer et al. | May 2006 | B2 |
7076391 | Pakzad et al. | Jul 2006 | B1 |
7082007 | Liu et al. | Jul 2006 | B2 |
7089355 | Auerbach et al. | Aug 2006 | B2 |
7113358 | Zayas et al. | Sep 2006 | B2 |
7120726 | Chen et al. | Oct 2006 | B2 |
7155448 | Winter | Dec 2006 | B2 |
7199981 | Zabtcioglu | Apr 2007 | B2 |
7254671 | Haswell | Aug 2007 | B2 |
7283316 | Chiao et al. | Oct 2007 | B2 |
7298568 | Ehrlich et al. | Nov 2007 | B2 |
7330323 | Singh et al. | Feb 2008 | B1 |
7343517 | Miller et al. | Mar 2008 | B2 |
7408731 | Uemura et al. | Aug 2008 | B2 |
7412585 | Uemura | Aug 2008 | B2 |
7436610 | Thelin | Oct 2008 | B1 |
7436614 | Uchida | Oct 2008 | B2 |
7440224 | Ehrlich et al. | Oct 2008 | B2 |
7486460 | Tsuchinaga et al. | Feb 2009 | B2 |
7490212 | Kasiraj et al. | Feb 2009 | B2 |
7509471 | Gorobets | Mar 2009 | B2 |
7516267 | Coulson et al. | Apr 2009 | B2 |
7529880 | Chung et al. | May 2009 | B2 |
7539924 | Vasquez et al. | May 2009 | B1 |
7603530 | Liikanen et al. | Oct 2009 | B1 |
7647544 | Masiewicz | Jan 2010 | B1 |
7669044 | Fitzgerald et al. | Feb 2010 | B2 |
7685360 | Brunnett et al. | Mar 2010 | B1 |
7840878 | Tang et al. | Nov 2010 | B1 |
7860836 | Natanzon et al. | Dec 2010 | B1 |
7885921 | Mahar et al. | Feb 2011 | B2 |
7900037 | Fallone et al. | Mar 2011 | B1 |
7982993 | Tsai et al. | Jul 2011 | B1 |
8006027 | Stevens et al. | Aug 2011 | B1 |
8031423 | Tsai et al. | Oct 2011 | B1 |
8116020 | Lee | Feb 2012 | B1 |
8179627 | Chang et al. | May 2012 | B2 |
8194340 | Boyle et al. | Jun 2012 | B1 |
8194341 | Boyle | Jun 2012 | B1 |
8341339 | Boyle et al. | Dec 2012 | B1 |
8443167 | Fallone et al. | May 2013 | B1 |
8560759 | Boyle et al. | Oct 2013 | B1 |
8687306 | Coker | Apr 2014 | B1 |
8693133 | Lee et al. | Apr 2014 | B1 |
8699185 | Teh | Apr 2014 | B1 |
20010042166 | Wilson et al. | Nov 2001 | A1 |
20030065872 | Edgar et al. | Apr 2003 | A1 |
20030220943 | Curran et al. | Nov 2003 | A1 |
20040019718 | Schauer et al. | Jan 2004 | A1 |
20040109376 | Lin | Jun 2004 | A1 |
20050069298 | Kasiraj et al. | Mar 2005 | A1 |
20050071537 | New et al. | Mar 2005 | A1 |
20050138265 | Nguyen et al. | Jun 2005 | A1 |
20050144517 | Zayas | Jun 2005 | A1 |
20050157416 | Ehrlich et al. | Jul 2005 | A1 |
20060090030 | Ijdens et al. | Apr 2006 | A1 |
20060112138 | Fenske et al. | May 2006 | A1 |
20060117161 | Venturi | Jun 2006 | A1 |
20060181993 | Blacquiere et al. | Aug 2006 | A1 |
20070016721 | Gay | Jan 2007 | A1 |
20070067603 | Yamamoto et al. | Mar 2007 | A1 |
20070174582 | Feldman | Jul 2007 | A1 |
20070204100 | Shin et al. | Aug 2007 | A1 |
20070226394 | Noble | Sep 2007 | A1 |
20070245064 | Liu | Oct 2007 | A1 |
20070288686 | Arcedera et al. | Dec 2007 | A1 |
20070294589 | Jarvis et al. | Dec 2007 | A1 |
20080098195 | Cheon et al. | Apr 2008 | A1 |
20080104308 | Mo et al. | May 2008 | A1 |
20080183955 | Yang et al. | Jul 2008 | A1 |
20080195801 | Cheon et al. | Aug 2008 | A1 |
20080256287 | Lee et al. | Oct 2008 | A1 |
20080256295 | Lambert et al. | Oct 2008 | A1 |
20080270680 | Chang | Oct 2008 | A1 |
20080307192 | Sinclair et al. | Dec 2008 | A1 |
20090019218 | Sinclair et al. | Jan 2009 | A1 |
20090043985 | Tuuk et al. | Feb 2009 | A1 |
20090055620 | Feldman et al. | Feb 2009 | A1 |
20090063548 | Rusher et al. | Mar 2009 | A1 |
20090119353 | Oh et al. | May 2009 | A1 |
20090150599 | Bennett | Jun 2009 | A1 |
20090154254 | Wong et al. | Jun 2009 | A1 |
20090164535 | Gandhi et al. | Jun 2009 | A1 |
20090164696 | Allen et al. | Jun 2009 | A1 |
20090187732 | Greiner et al. | Jul 2009 | A1 |
20090193184 | Yu et al. | Jul 2009 | A1 |
20090198952 | Khmelnitsky et al. | Aug 2009 | A1 |
20090204750 | Estakhri et al. | Aug 2009 | A1 |
20090222643 | Chu | Sep 2009 | A1 |
20090240873 | Yu et al. | Sep 2009 | A1 |
20090271581 | Hinrichs, Jr. | Oct 2009 | A1 |
20090276604 | Baird et al. | Nov 2009 | A1 |
20100011275 | Yang | Jan 2010 | A1 |
20100061150 | Wu et al. | Mar 2010 | A1 |
20100161881 | Nagadomi et al. | Jun 2010 | A1 |
20100169543 | Edgington et al. | Jul 2010 | A1 |
20100169551 | Yano et al. | Jul 2010 | A1 |
20100208385 | Toukairin | Aug 2010 | A1 |
20110167049 | Ron | Jul 2011 | A1 |
20110304935 | Chang et al. | Dec 2011 | A1 |
Number | Date | Country |
---|---|---|
2009102425 | Aug 2009 | WO |
Entry |
---|
Amer, Ahmed et al. (May 2010) “Design Issue for a Shingled Write Disk System” 26th IEEE Symposium on Massive Storage Systems and Technologies: Research Track. |
Rosenblum, (Jun. 1992) “The Design and Implementation of a Log-structured File System”, EECS Department, University of California, Berkeley, Technical Report No. UCB/CSD-92-696, pp. 1-93. |
Rosenblum, Mendel and Ousterhout, John K. (Feb. 1992), “The Design and Implementation of a Log-Structured File System.” ACM Transactions on Computer Systems, vol. 10, Issue 1, pp. 26-52. |
Definition of adjacent, Merriam-Webster Dictionary, retrieved from http://www.merriam-webster.com/dictionary/adjacent on Oct. 30, 2013 (1 page). |
Re:Hard drive Inner or Outer tracks???, Matthias99, Apr. 12, 2004, retrieved from http://forums.anandtech.com/showthread.php?p=11 055300 on Oct. 29, 2013. |
You Don't Know Jack about Disks, Dave Anderson, Seagate Technologies, Queue—Storage Queue, vol. 1, issue 4, Jun. 2003, pp. 20-30 (11 pages). |