Faster write operations to nonvolatile memory by manipulation of frequently accessed sectors

0001The present invention relates generally to a method and apparatus for decreasing the speed of write operations in systems utilizing nonvolatile memory and particularly to decreasing the speed of write operations in such systems when performing random write operations. Publications by the inventor in related subject matter include <patcit id="pcit0001" dnum="US6145051A"><text>US-A-6,145,051</text></patcit>, <patcit id="pcit0002" dnum="US5907856A"><text>US-A-5,907,856</text></patcit> and <patcit id="pcit0003" dnum="US5845313A"><text>US-A-5,845,313</text></patcit>.

0002With the advent of the digital information revolution, nonvolatile memory (or FLASH memory or EEPROM memory) has enjoyed considerable popularity within less than a decade. This is in large part due to the particular characteristics, known to those of ordinary skill in the art, attributed to nonvolatile memory, such as maintenance of information even when power is interrupted or disconnected.

0003Nonvolatile memory has a wide variety of utilization, some of which include digital cameras where they are used to store photographs, in digital format, for later editing or loading elsewhere and personal computers (PCs) where they are used to store various types of digital information.

0004One of the problems posing a constant challenge to those interested in using nonvolatile memory is known to be the speed of read and write operations. As after each write operation, which involves changing the state of only cells that are still at a logical one state, an erase operation is necessary prior to a subsequent write operation to the same location, it is clear that any improvements to the speed of a write operation proves to be an impressive improvement to the overall system performance.

0005By way of a brief background to the way in which information is organized within nonvolatile memory, the memory, which may be one or more integrated circuits, typically more than one, is prearranged into blocks with each block including a predetermined number of sectors and each sector including a predetermined number of bytes. A sector is identified by the system by a logical block address (LBA) but when it is to be written to or read from nonvolatile memory, it is identified by a physical block address (PBA). There is generally a known correspondence between LBAs and PBAs, which may be maintained in random access memory (RAM).

0006An individual block or an individual sector may be defined as an erasable unit depending on the user. For example, a sector might and typically does include 512 bytes of data and 16 bytes of overhead information and in a given application, during an erase operation, the entire sector may be erased, which include many cells. In fact, a number of sectors may be erased at a time, in which case, the user has probably designed the system so that one or more blocks are erased together.

0007In existing systems employing nonvolatile memory, two types of data are stored therein. One type is file system data, another type is user system data. File system data provides information to the operating system as to the location in which a file, which includes user data, is located. User system data is the contents of a file, each file is generally comprised of large blocks of data typically requiring sequential write operations. In sequential write operations, LBAs appear in sequential order, i.e. sequential sectors are being written thereto.

0008Write operations are generally initiated by a host that interacts with a controller to direct information in and out of the memory. Sectors, identified by the host, are written thereto in memory. In mass storage applications, the host may initiate two types of write operation, one is a sequential write operation, as discussed hereinabove, and the other is a random write operation. One method of random write operation is generally performed when updating file system data.

0009In the more recent Windows operating system, employed in most PCs and workstations, a new file system, referred to as "FAT32" (File Allocation Table 32) is utilized, which allows file systems of the size greater than 2Gigabytes. Advantages allowed by using FAT32 include formatting cards used for various applications, such as digital cameras, that have much larger capacity than prior file systems would have allowed.

0010In FAT32 type file systems, a special sector, referred to as "FSInfo" (File System Information) is employed. The FSInfo sector is continuously updated with information regarding the location of the next free cluster. In effect, this saves the file system software (operating system) from having to search the FAT for the location of a free cluster, which can be a time-consuming task. Not only FSInfo sectors but certain types of sectors as well, are generally written or re-written thereto more often than types of sectors.

0011However, continuous writing or updating of these certain sectors that are written thereto more often than others has the side effect of slowing down the system performance, in general, as well as consuming much nonvolatile memory space thereby causing inefficiencies in the use of nonvolatile memory.

0012As stated above, nonvolatile memory is organized into sectors with a group of sectors comprising a block. A sector typically includes 512 bytes of user data and 16 bytes of overhead information and a block may include 256 sectors, although, a sector may have a different number of bytes and a block may have a different number of sectors.

0013In existing nonvolatile memory systems, when a sector is to be written thereto, a free block is first located. A logical block address (LBA) value is correlated with a new physical block address (PBA) addressing a free or available block in nonvolatile memory. At this point, the free found block is considered to be 'open' or 'pending', i.e. being written thereto with sector information. For the purposes of discussion, the free found block, having an 'open' or 'pending' status, will be referred to as block 0, identified by VLBA 0.

0014Next, a sector subsequent to that being written is examined and if it is found that the subsequent sector is sequential or belongs to the correlated VLBA block, i.e. block 0 (PBA 0), the sector information is written to the sector location within the same block. Succeeding sectors of information, which belong to the correlated VLBA block, i.e. block 0, continue to be written thereto until block 0 is no longer free and has no more free locations in which case a different free block, for example, block 10 identified by VLBA 10 (PBA 10), is designated to be written thereto. Now, if there are further writes or updates of an already written sector of block 0, they will be written or stored in block 10. In this case, all of the sector information that was not updated is moved from block 0 to block 10 and block 0 is 'closed' pending erasure thereof.

0015In the previous example, after block 0 is no longer free, and assuming the new write command has sector information, which belongs to the same VLBA block, i.e. block 0, and is now being updated, a different free block, for example, block 10, identified by VLBA 0 and VPBA 10, is designated to be written thereto and is thus used to store updated sector information. At this point, blocks 0 and 10 are 'pending' or 'open'. If yet another write command is received commanding updating of sector information that does not belong to block 0 and thus corresponds to an LBA other than VLBA 0, yet another free block, i.e. block 20 identified by VLBA 1 and VPBA 20 is used to store such update. An example of sector information that does not belong to the same VLBA block would be non-sequential sectors relative to those sectors being stored in block 0 or block 10. In this example, if block 0 and therefore block 10 are each designated to store sectors 0-255 (LBA 0 ... LBA 255), then a write to sector 614 (LBA 614) is considered a write to a sector that does not belong to the same VLBA block as that of block 0 or block 10. All of the sector information that was not updated however is moved from block 0 to block 10. Now, in this case, blocks 10 and 20 will be 'open' or 'pending' and block 0 will be 'closed'. As explained in <patcit id="pcit0004" dnum="US620544A" dnum-type="L"><text>U.S. patent application, Serial No. 09/620,544</text></patcit>, entitled "Moving Sectors Within A Block of Information In A Flash Memory Mass Storage Architecture" and filed on July 21, 2000, and <patcit id="pcit0005" dnum="US5845313A"><text>U.S. Patent No. 5,845,313</text></patcit>, entitled "Direct Logical Block Addressing Flash Memory Mass Storage Architecture," and issued on December 1, 1998.

0016Still using the same example as above, yet another scenario is the case where even prior to block 0 running out of free space, that is, while it remains free or has available sector locations, if the new write command is a write operation to one of the sectors that was already updated once, again, another block, such as block 10, is used to store the second update of the already once-updated sector information and subsequent sectors, to the extent they are in sequential order and/or belong to the same LBA group as that of block 0, are written to corresponding sector locations of block 10. Un-updated sectors of block 0 are eventually moved to block 10 and block 0 is 'closed'. Block 10 remains 'open' or 'pending'.

0017To this end, in the case where a block includes, for example, 256 sectors, a maximum of possible 255 move operations of sector information would have to be performed, which is quite time consuming and degrades system performance.

0018This is illustrated, in high-level block diagram form, in <figref idref="f0001">Fig. 1</figref>, where a nonvolatile memory system 10 is shown to include a block 12 for storing sector information. The block 12 is initially free and subsequently filled with sector information if and to the extent to which these sectors are sequential. Once a non-sequential sector is to be written, the sectors, with block 12, which include recent or 'good' information, are moved to the block 14, as shown by the arrow 16, which could be 255 move operations (read of 255 sectors from the old or previous block and write of the same to the new block). The move operation of the good or un-updated sectors of the previous block to the new block, are known to cause a 50% slower system performance in nonvolatile memory system.

0019Thus, the need arises for a system and method for minimizing the time required for performing writing operations of non-sequential sectors in the file system area of nonvolatile memory.

0020<patcit id="pcit0006" dnum="US5835935A"><text>US-A-5835935</text></patcit> describes a memory system in which memory is partitioned into user files and system files, the system files being subdivided into clusters each comprising a plurality of sectors. Each cluster stores the system file for a single predetermined LBA and, as information within the LBA is changed, the new information is written into an empty sector within the cluster. Once filled, the cluster is erased for recycling and an empty cluster located to repeat the process.

0021Aspects of the present invention are set out in the appended independent claims.

0022Briefly, an embodiment of the present invention includes a digital equipment system having a host for sending random write commands to write files having sector information and having a controller device responsive to the commands for writing and updating FSInfo sector information. The controller controls a nonvolatile memory system (an example of nonvolatile memory is flash memory, as well known to those of ordinary skill in the art) organized into blocks, each block including a plurality of sector locations for storing sector information, a particular free block, designated for storing FSInfo sector information. Upon updating of the FSInfo sector, the updated FSInfo sector information is written to a next free sector of the dedicated block thereby avoiding moving the sectors of the particular block to another block, hence, improving system performance.

0023The foregoing and other objects, features and advantages of the present invention will be apparent from the following detailed description of the preferred embodiments which make reference to several figures of the drawing.

IN THE DRAWINGS

0024<ul id="ul0001" list-style="none"><li><figref idref="f0001">Fig. 1</figref> shows a prior art nonvolatile memory system 10.</li><li><figref idref="f0001">Fig. 2</figref> illustrates a digital equipment system 50 in accordance with an embodiment of the present invention.</li><li><figref idref="f0002">Fig. 2(a)</figref> shows another embodiment of the present invention.</li><li><figref idref="f0002">Fig. 2(b)</figref> shows the management of frequently accessed sectors by the nonvolatile memory system 20.</li><li><figref idref="f0003">Fig. 3</figref> shows a high-level block diagram of a digital system 500, such as a digital camera, employing the nonvolatile memory system 20 of <figref idref="f0001">Fig. 2</figref>, in accordance with an embodiment of the present invention.</li></ul>

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

0025Referring now to <figref idref="f0001">Fig. 2</figref>, a nonvolatile memory system 20 is shown to include a previous block 22 and a free block 24, in accordance with an embodiment of the present invention. The nonvolatile memory system 20 is comprised of nonvolatile memory storage locations that are organized into blocks with each block including a group of sector locations for storing sector information. Sector information is generally comprised of user data and overhead data. User data is data that is intended to be stored by a user, such as photographs, music and the like, and overhead data is information relating to the user data, such as its location and/or error information and so forth. Various types of sector information are known to those of ordinary skill in the art, one of which is FSInfo sector.

0026The previous block 22 is dedicated to storing a particular type of sector, namely, a sector that is accessed more frequently than others. This type of sector is detected by using a predetermined threshold for the number of times a sector is accessed and if the number of times a sector is accessed exceeds the threshold, then, it will be treated differently, as discussed hereinbelow. The threshold value is programmed as a fixed value and used by the firmware or software that is directing the operation of the nonvolatile memory system, within the controller. In one embodiment of the present invention, an access counter is used to maintain track of the number of times a sector is accessed.

0027An example of an access counter value 21 is shown in <figref idref="f0001">Fig. 2</figref>, within the overhead portion 25 the first sector of the block 22. This sector is also shown to include a user data portion 23. The access counter is physically located within the controller controlling the nonvolatile memory, an example of which is shown by the access counter 521, in <figref idref="f0003">Fig. 3</figref>, shown coupled to the flash controller block 532. Alternatively, the access counter may be located outside of the controller.

0028Another way of identifying these types of sectors is by a comparison of the number of times they are accessed relative to each other and if sectors are detected that are in the category of being accessed more so than others, they are treated special, as noted below.

0029In both of these methods, a counter, i.e. the access counter, is used to count each time a sector is accessed; the counter value for each sector is generally stored in the overhead portion of the sector information in nonvolatile memory in one example embodiment.

0030In one example, such as the example of FAT 32, system traffic is analyzed and a sector that is accessed more often than other sectors, such as the FSInfo sector is detected as a frequently-accessed sector,or if the sector, which belongs to a VLBA block and is accessed repeatedly during writing a file. The FSInfo sector is an example of such a sector. While the discussion below is mostly regarding the FSInfo sector, it should be noted that the present invention applies to whatever kind of sector is accessed more frequently than others and is not limited to the FSInfo sector and that the latter is merely an example embodiment of the present invention. While a discussion of the FSInfo sector is provided hereinabove, a part of the same, namely the function of the FSInfo sector, is repeated.

0031In FAT32 type file systems, a special sector, referred to as "FSInfo" is employed. The FSInfo sector is continuously updated with information regarding the location of the last known free cluster. In effect, this saves the file system software (operating system) from having to search the FAT for the location of a free cluster, which can be a time-consuming task.

0032As noted above, FSInfo sector is merely an example of a sector that is frequently accessed. The present invention applies to any type of sector that is detected or known to be accessed more frequently than other sectors and that may cause an unnecessary move if all of the data is not being updated and belongs to the same VLBA block. At any given time, one of the locations of the block 22 stores a current version of the FSInfo sector. For example, the first version of the FSInfo sector might be stored at location 26 of the block 22 and when the FSInfo sector is updated, the updated version may be stored at location 28 and still the following version of the FSInfo sector may be stored at location 30 of the block 22 and so on and so forth until all of the locations of the block 22 are written thereto with various versions of the FSInfo sector with the most recent version being stored at location 'n' of block 22 with 'n' being an integer, such as 256 in the case where there are 256 sectors per block. Thus, every time the FSInfo sector is updated, it is written to a new (or next although it need not be the next location) sector location within the block 22 and the previous sector location is designated as being 'old'. This designation is so that upon a read operation, the most recent version of the FSInfo sector can be easily identified. That is, the sector location, which includes the most recent version of the FSInfo sector will have a designation other than 'old', whereas, all sector locations, which include a previous version of the FSInfo sector will have a designation of 'old'. There are other ways of identifying the most recent version of FSInfo sector, one way is to keep track of the LBA value identifying the sector location, which includes the most recent version of FSInfo sector.

0033Once the block 22 is full, that is, all of its sector locations are written thereto with various versions of the FSInfo sector, the block 24 is written thereto and the block 22 is ready to be erased and an 'old' designation of the same is made. Again, with respect to block 24, as with block 22, the FSInfo sector is written to a sector location thereof each time the FSInfo sector is updated. To reiterate, the updating of the FSInfo sector in prior art techniques, as noted hereinabove, required moving the 'good' sectors of the previous block to another block (or read and written) resulting in time-consuming operations, i.e. many read and write operations, that need not occur in the present invention.

0034In the present example, there is no moving of all of the 'good' sectors of a block to another block when the FSInfo sector is updated, accordingly, there is no need to move, for example, 255 sectors from the block 22 to the block 24, as there would have been in prior art systems.

0035In the example of <figref idref="f0001">Fig. 2</figref>, the next free sector location within the same block 22 is used to store the updated or most recent version of FSInfo sector rather than a sector within a new block.

0036In the present example, each time FSInfo sector is updated, a write operation occurs. In the present example, there is no requirement to move, i.e. read and write, the sectors of the block 22 to the block 24. The block 22 is dedicated to storing no other type of sectors other than the FSInfo sector. Upon completion of writing to all of the sector locations of the block 22 and upon receiving the next update of the FSInfo sector and storing the same in a new block, the block 22 is erased.

0037In an embodiment of the present invention, more than one sector is treated differently, in that, there are more than one type of sectors that are accessed much more often than others and are thus dedicated, at all times, a block. This is perhaps best understood using an example.

0038Assuming that sectors, identified by LBA 0 and LBA 50, are accessed much more frequently than others, every time LBA 0 is written, it will be written to a particular block within the nonvolatile memory and following the same, when LBA 50 is written, it is also written to the particular block and then again, when LBA 0 is re-written, rather than being written to a different block, as done by the prior art, it is written to the particular block (perhaps the third location, as the first two locations are occupied by the first writes to the LBA 0 and LBA 50). Next, upon a re-write of sector 50, it too is written to the particular block and so forth and so on until all of the sector locations of the particular block are written thereto. Afterwards, a different block is dedicated to which following re-writes of LBA 0 and LBA 50 are made and the particular block is erased.

0039An example is shown in <figref idref="f0002">Fig. 2(a)</figref> where blocks 44, 46 and 48 are designated as blocks accommodating special sectors that are accessed more frequently than others. In this case, as an example, sector information identified by LBA 0 is accessed frequently and every time it is written, it is written to a different block. The first time it is being written, it is written to the sector location 50 of block 44, the second time it is being written, it is written to the location 52 of block 46 and so on and so forth, such as another writing of the same to location 54 of block 48. This avoids any move operations and thereby increases system performance.

0040The drawback is that if done with many blocks, i.e. the number of blocks such as 44, 46 and 48 to which the LBA 0 sector information is written is numerous, many blocks are left 'open' or are written thereto and thus can not be used to write other kinds of information until at some point, they are 'closed' or declared of no further use and erased before being used again.

0041<figref idref="f0002">Fig. 2(b)</figref> shows the management of frequently accessed sectors by the nonvolatile memory system 20. The system 20, in <figref idref="f0002">Fig. 2(b)</figref>, is shown to include a block 0, a block 1 and a block M, while many blocks may be employed, blocks 0, 1 and M are shown for illustrative purposes.

0042Each block includes a predetermined number of sector storage locations, such as 16, each for storing sector information, i.e. user data and overhead information. Block 0 is identified or addressed by VPBA 0, block 1 is identified by VPBA 1 and block M is identified by VPBA 17, by way of example. Each of the sector storage locations is identified by a particular PBA. For example, the first sector storage location 60 of block 0 is identified by a PBA 0 and the last sector storage location 62 of block 0 is identified or addressed by PBA 15. While not shown, the remaining blocks of <figref idref="f0002">Fig. 2(b)</figref> include similar PBA addressing of their respective sector storage locations.

0043In the example of <figref idref="f0002">Fig. 2(b)</figref>, LBA 15 is chosen for frequently accessing thereof by the host and will thus be the main target of this discussion. It should be noted that any other sector may be designated as a frequently-accessed sector, as discussed hereinabove. Block M serves as a dedicated block for storing updates to the sector identified by LBA 15, referred to at times, as sector 15.

0044First, sectors are written to the sector storage locations of block 0. Upon a re-write of these sectors, they are written to the sectors of block 1 with the exception of sector 15, or that sector identified by LBA 15. That is, rather than writing sector 15 to block 1, it is written to the dedicated block, block M, in this case, as are subsequent updates to sector 15. Thus, everytime sector 15 is updated either as part of a sequential number of sectors, i.e. sectors 0-15, or otherwise, such as in a random write operation, it is written to an unused or available sector location within the block M. The first time it is updated, it may be written to the sector location 66, the next time it is updated it may be written to the sector location 68 and so on until it is written to the sector location 70. Thereafter, another available block is allocated to include updates of sector 15.

0045<figref idref="f0003">Fig. 3</figref> depicts a digital system 500 such as a digital camera, personal computer, or the like, employing an embodiment of the present invention. Digital system 500 is illustrated to include a host 502, which may be a personal computer (PC) or simply a processor of any generic type commonly employed in digital systems, coupled to a controller circuit 506 for storing in and retrieving information from nonvolatile memory unit 508. In one embodiment of the present invention, the memory unit 508 includes the nonvolatile memory system 20 of <figref idref="f0001">Fig. 2</figref> and is managed similarly to that which is described with respect to preceding figures hereinabove.

0046The controller circuit 506 may be a semiconductor (otherwise referred to as an "integrated circuit" or "chip") or optionally a combination of various electronic components. In the preferred embodiment, the controller circuit is depicted as a single chip device. The nonvolatile memory unit 508 is comprised of one or more memory devices, which may each, be flash or EEPROM types of memory. In the preferred embodiment of <figref idref="f0003">Fig. 3</figref>, memory unit 508 includes a plurality of flash memory devices, 510 - 512, each flash device includes individually addressable locations for storing information. In the preferred application of the embodiment in <figref idref="f0003">Fig. 3</figref>, such information is organized in blocks with each block having one or more sectors of data. In addition to the data, the information being stored may further include status information regarding the data blocks, such as flag fields, address information and the like.

0047The host 502 is coupled through host information signals 504 to a controller circuit 506. The host information signals comprise of address and data busses and control signals for communicating command, data and other types of information to the controller circuit 506, which in turn stores such information in memory unit 508 through flash address bus 512, flash data bus 514, flash signals 516 and flash status signals 518 (508 and 512-516 collectively referred to as signals 538). The signals 538 may provide command, data and status information between the controller 506 and the memory unit 508.

0048The controller 506 is shown to include high-level functional blocks such as a host interface block 520, a buffer RAM block 522, a flash controller block 532, a microprocessor block 524, a microprocessor controller block 528, a microprocessor storage block 530, a microprocessor ROM block 534, an ECC logic block 540 and a space manager block 544. The host interface block 520 receives host information signals 504 for providing data and status information from buffer RAM block 522 and microprocessor block 524 to the host 502 through host information signals 504. The host interface block 520 is coupled to the microprocessor block 524 through the microprocessor information signals 526, which is comprised of an address bus, a data bus and control signals.

0049The microprocessor block 524 is shown coupled to a microprocessor controller block 528, a microprocessor storage block 530 and a microprocessor ROM block 534, and serves to direct operations of the various functional blocks shown in <figref idref="f0003">Fig. 3</figref> within the controller 506 by executing program instructions stored in the microprocessor storage block 530 and the microprocessor ROM block 534. Microprocessor 524 may, at times, execute program instructions (or code) from microprocessor ROM block 534, which is a nonvolatile storage area. On the other hand, microprocessor storage block 530 may be either volatile, i.e., read-and-write memory (RAM), or nonvolatile, i.e., EEPROM, type of memory storage. The instructions executed by the microprocessor block 524, collectively referred to as program code, are stored in the storage block 530 at some time prior to the beginning of the operation of the system of the present invention. Initially, and prior to the execution of program code from the microprocessor storage location 530, the program code may be stored in the memory unit 508 and later downloaded to the storage block 530 through the signals 538. During this initialization, the microprocessor block 524 can execute instructions from the ROM block 534.

0050Controller 506 further includes a flash controller block 532 coupled to the microprocessor block 524 through the microprocessor information signals 526 for providing and receiving information from and to the memory unit under the direction of the microprocessor. Information such as data may be provided from flash controller block 532 to the buffer RAM block 522 for storage (may be only temporary storage) therein through the microprocessor signals 526. Similarly, through the microprocessor signals 526, data may be retrieved from the buffer RAM block 522 by the flash controller block 532.

0051ECC logic block 540 is coupled to buffer RAM block 522 through signals 542 and further coupled to the microprocessor block 524 through microprocessor signals 526. ECC logic block 540 includes circuitry for generally performing error coding and correction functions. It should be understood by those skilled in the art that various ECC apparatus and algorithms are commercially available and may be employed to perform the functions required of ECC logic block 540. Briefly, these functions include appending code that is for all intensive purposes uniquely generated from a polynomial to the data being transmitted and when data is received, using the same polynomial to generate another code from the received data for detecting and potentially correcting a predetermined number of errors that may have corrupted the data.

0052ECC logic block 540 performs error detection and/or correction operations on data stored in the memory unit 508 or data received from the host 502.

0053The space manager block 544 employs a preferred apparatus and algorithm for finding the next unused (or free) storage block within one of the flash memory devices for storing a block of information, as will be further explained herein with reference to other figures. As earlier discussed, the address of a block within one of the flash memory devices is referred to as VPBA, which is determined by the space manager by performing a translation on an LBA received from the host. A variety of apparatus and method may be employed for accomplishing this translation. An example of such a scheme is disclosed in <patcit id="pcit0007" dnum="US5845313A"><text>U.S. Pat. No. 5,845,313</text></patcit>, entitled "Direct Logical Block Addressing Flash Memory Mass Storage Architecture".

0054Other LBA to PBA translation methods and apparatus may be likewise employed.

0055Space manager block 544 includes SPM RAM block 548 and SPM control block 546, the latter two blocks being coupled together. The SPM RAM block 548 stores the LBA-PBA mapping information (otherwise herein referred to as translation table, mapping table, mapping information, or table) under the control of SPM control block 546. This mapping can be kept also in the non-volatile memory. Alternatively, the SPM RAM block 548 may be located outside of the controller, such as shown in <figref idref="f0003">Fig. 3</figref> with respect to RAM array 100.

0056The RAM memory is arranged to be addressable by the same address as the LBA provided by the host. Each such addressable location in the RAM includes a field which holds the physical address of the data in the nonvolatile mass storage expected by the host.

0057Ultimately, the way in which a block is addressed in nonvolatile memory is by a virtual logical block address (VLBA), which is a modified LBA, for each block. The way in which a sector is addressed within each block is by the use of a virtual physical block address (VPBA). The VPBA locations are for storing information generally representing a PBA value corresponding to a particular LBA value.

0058In operation, the host 502 writes and reads information from and to the memory unit 508 during for example, the performance of a read or write operation through the controller 506. In so doing, the host 502 provides an LBA to the controller 506 through the host signals 504. The LBA is received by the host interface block 520. Under the direction of the microprocessor block 524, the LBA is ultimately provided to the space manager block 544 for translation to a PBA and storage thereof, as will be discussed in further detail later.

0059Under the direction of the microprocessor block 524, data and other information are written into or read from a storage area, identified by the PBA, within one of the flash memory devices 510-512 through the flash controller block 532. The information stored within the flash memory devices may not be overwritten with new information without first being erased, as earlier discussed. On the other hand, erasure of a block of information (every time prior to being written), is a very time and power consuming measure. This is sometimes referred to as erase-before-write operation. The preferred embodiment avoids such an operation by continuously, yet efficiently, moving a sector (or multiple sectors) of information, within a block, that is being re-written from a PBA location within the flash memory to an unused PBA location within the memory unit 508 thereby avoiding frequent erasure operations. A block of information may be comprised of more than one sector such as 16 or 32 sectors. A block of information is further defined to be an individually-erasable unit of information. In the past, prior art systems have moved a block stored within flash memory devices that has been previously written into a free (or unused) location within the flash memory devices. Such systems however, moved an entire block even when only one sector of information within that block was being re-written. In other words, there is waste of both storage capacity within the flash memory as well as waste of time in moving an entire block's contents when less than the total number of sectors within the block are being re-written. The preferred embodiments of the present invention, as discussed herein, allow for "moves" of less than a block of information thereby decreasing the number of move operations of previously-written sectors, consequently, decreasing the number of erase operations.

0060Although the present invention has been described in terms of specific embodiments it is anticipated that alterations and modifications thereof will no doubt become apparent to those skilled in the art.