Patent Application
20060143395
The present invention generally relates to the field of digital memory storage. More specifically, the present invention relates to a method and apparatus for managing data in a cache memory of a mass-storage device and/or system.
A cache memory is a high-speed buffer located between a host(s) and a disk device(s), which is meant to reduce the overall latency of Input/Output (I/O) activity between the storage system and a host accessing data on the storage system. Whenever a host requests data stored in a memory system, the request may be served with significantly lower latency if the requested data is already found in cache, since this data must not be brought from the disks. As of the year 2004, speeds of IO transactions involving disk activity are typically on the order of 5-10 milliseconds, whereas IO speeds involving cache (e.g. RAM memory) access are on the order of several nanoseconds.
The relatively high latency associated with disk activity derives from the mechanical nature of the disk devices. In order to retrieve requested data from a disk based device, a disk controller must first cause a disk reading arm to physically move to a track containing the requested data. Once the head of the arm has been placed at the beginning of a track containing the data, the time required to read the accessed data on the relevant track is relatively very short, on the order of several microseconds.
One criteria or parameter which is often used to measure the efficiency of a cache memory system or implementation is a criteria referred to as a “hit ratio”. A hit ratio of a specific implementation is the percentage of “data read” requests issued by the host that are already found in cache and that consequently did not require time intensive retrieval from disk operations. An ideal cache system would be one reaching a 100% hit ratio.
Intelligent cache algorithms aimed at raising the level of hit ratio to a maximum often implement in the cache (amongst other things) replacement procedures that establish a “discard priority”. The discard priority relates to data portions which are already stored in the cache, and is responsible for managing the order by which each of the data portions that are currently stored in the cache are to be discarded. The various approaches to the management of the discard priority queue attempt to anticipate which of the data portions currently stored in the cache are the least likely to be requested by the host, and set the priority queue accordingly. Thus, whenever cache space is required for fetching new data into the cache, the discard priority queue is consulted, and the data portions which have been determined to be the least likely to be requested by the host are discarded first, in order to free up cache storage space which may be used for storing the newly fetched data. From another perspective, the various approaches to the management of the discard priority queue are intended to increase the likelihood that those portions of data that have been determined to be the most likely to be requested by the host will not be discarded prematurely in order to free up space or for any other purpose.
The link between certain data portions in the cache is significant for the purposes of replacement procedures. When a certain data portion has been addressed by a host as part of a certain group of operatively linked data portions or as part of a certain file, for example, it is likely that when that data portion is next addressed it will be addressed as part of the group or as part of the file, rather than as an isolated data portion. Furthermore, while the usage of the cache's memory resources grows with the amount of data fetched into the cache, fetching-time does not increase with the amount of data to be fetched into the cache. Thus, while fetching a large group and a small group (e.g., one) of linked data portions will require substantially the same amount time, discarding the larger group of linked data will free-up substantially more storage space.
Accordingly, it would be beneficial to manage the cache in terms of groups of operatively linked data portions, rather than in terms of discrete data portions. Furthermore, it would be beneficial to assign a discard priority to such groups of operatively linked data portions at least in accordance with the size each group. Yet furthermore, it would be beneficial to assign a discard priority to such groups of operatively linked data portions, such that, all other things being equal, larger groups are assigned with a higher priority to be discarded.
In accordance with some embodiments of the present invention, there is provided a method of managing a cache memory in a mass-storage system and a cache management module for managing a cache memory device. In accordance with some embodiments of the present invention, the cache management module may include a groups management module and a replacement procedure module. In accordance with some embodiments of the present invention, the groups management module may be adapted to define groups of allocation units in accordance with at least an operative criterion and to create a new group of allocation units whenever it is determined that in accordance with at least the operative criterion none of the one or more existing groups is appropriate to include an allocation unit, and the replacement procedure module may be adapted to manage the cache in terms of groups of allocation units, rather than in terms of discrete allocation units.
According to further embodiments of the present invention, in accordance with the operative criterion, the groups management module may be adapted to add the allocation unit to an existing group of allocation units in case the data associated with the allocation unit is in sequence with the data associated with the group of allocation units. According to yet further embodiments of the present invention, the groups management module may be adapted to add the allocation unit to an existing group of allocation units in case, in addition to being in sequence with the data associated with the allocation unit, the data associated with the group of allocation units was accessed while being stored in the cache.
In accordance with some embodiments of the present invention, the replacement procedure module may be adapted to create two or more queues, each being associated with groups of allocation units including a predefined number of allocation units, and to assign each of the groups of allocation units to one of the two or more queues in accordance with the size of the group of allocation units. According to further embodiments of the present invention, the replacement procedure module may be further adapted to assign to each of one or more of the replacement queues a predefined aging factor. The replacement procedure module may be further adapted to modify the age of one or more the groups or of one or more allocation units belonging to one or more groups in accordance with the aging factor associated with the replacement queue to which the groups have been assigned.
In accordance with some further embodiments of the present invention, the weight of the aging factor associated with each of the one or more queues may be associated with the size of the groups with which the queue is associated. In accordance with further embodiments of the present invention, the weight of the aging factors may be selected in a manner to cause the discarding of larger groups of allocation units to be accelerated. The replacement procedure module may be adapted to select the group of allocation units to be discarded in accordance with the modified age of one or more of the groups. For example, according to some embodiments of the present invention, the replacement procedure module may be adapted to select the allocation units which are associated with the oldest modified age.
Further embodiments of the present invention relate to a method of managing a cache memory in a mass-storage system. According to some embodiments of the method in accordance with the present invention, groups of allocation units may be defined in accordance with at least an operative criterion, such that a new group of allocation units is to be created whenever it is determined that in accordance with at least the operative criterion none of the one or more existing groups is appropriate to include an allocation unit, and the cache may be managed in terms of groups of allocation units, rather than in terms of discrete allocation units.
In order to understand the invention and to see how it may be carried out in practice, a preferred embodiment will now be described, by way of non-limiting example only, with reference to the accompanying drawings, in which:
It will be appreciated that for simplicity and clarity of illustration, elements shown in the figures have not necessarily been drawn to scale. For example, the dimensions of some of the elements may be exaggerated relative to other elements for clarity. Further, where considered appropriate, reference numerals may be repeated among the figures to indicate corresponding or analogous elements.
In the following detailed description, numerous specific details are set forth in order to provide a thorough understanding of the invention. However, it will be understood by those skilled in the art that the present invention may be practiced without these specific details. In other instances, well-known methods, procedures, components and circuits have not been described in detail so as not to obscure the present invention.
Unless specifically stated otherwise, as apparent from the following discussions, it is appreciated that throughout the specification discussions utilizing terms such as “processing”, “computing”, “calculating”, “determining”, or the like, refer to the action and/or processes of a computer or computing system, or similar electronic computing device, that manipulate and/or transform data represented as physical, such as electronic, quantities within the computing system's registers and/or memories into other data similarly represented as physical quantities within the computing system's memories, registers or other such information storage, transmission or display devices.
Embodiments of the present invention may include apparatuses for performing the operations herein. This apparatus may be specially constructed for the desired purposes, or it may comprise a general purpose computer selectively activated or reconfigured by a computer program stored in the computer. Such a computer program may be stored in a computer readable storage medium, such as, but is not limited to, any type of disk including floppy disks, optical disks, CD-ROMs, magnetic-optical disks, read-only memories (ROMs), random access memories (RAMs) electrically programmable read-only memories (EPROMs), electrically erasable and programmable read only memories (EEPROMs), magnetic or optical cards, or any other type of media suitable for storing electronic instructions, and capable of being coupled to a computer system bus.
The processes and displays presented herein are not inherently related to any particular computer or other apparatus. Various general purpose systems may be used with programs in accordance with the teachings herein, or it may prove convenient to construct a more specialized apparatus to perform the desired method. The desired structure for a variety of these systems will appear from the description below. In addition, embodiments of the present invention are not described with reference to any particular programming language. It will be appreciated that a variety of programming languages may be used to implement the teachings of the inventions as described herein.
Throughout the specification and the claims, the term “allocation unit” or the like, shall be used to describe an allocatable unit of cache storage space which may be allocated by a cache or by a cache's controller for storing data within the allocated storage space.
The terms “consecutive”, “successive”, “following”, “preceding”, “in sequence with” and the like shall be used throughout the specification and the claims to describe a corresponding relation between the respective indices of two or more allocation units in a cache storage memory. It should be noted, that an allocation unit as used herein may be comprised of one or more (commonly successive) data blocks. In the field of mass-storage systems, and other systems, one or more units of data are defined in the system and are indexed for purposes of data management, data traffic, data transfer, scaling, etc. Thus each data unit in the system is associated with one or more indicia. For example, in Storage Area Networks (SAN), several data units may be defined and indexed including the following: Logical Unit or Logical Volume (LU)—a LU is a string of consecutive blocks indicated by logical block addresses (LBAs) starting from 0 and going up to N, where N is the size of the LU; Partitions—a partition is a subunit of a LU and is comprised of a string of consecutive blocks which correspond to a certain segment of the LU of which the partition is part of. Partitions are indexed units and are commonly defined for purposes of internal management of data in the system. Thus, a LU consists of consecutive partitions; Allocation units—an allocation unit, as mentioned above, relates to an allocatable unit of cache storage space which may be allocated by a cache or by a cache's controller for storing data within the allocated storage space. In some storage systems, each allocation unit in the cache corresponds to one (whole) partition of the storage system. In other storage systems, such as the one described in U.S. Provisional Patent Application No. 60/639,780, filed Dec. 29, 2004 and in U.S. patent application Ser. No. 11/123,633 filed, May 6, 2005, allocation units are subunits of a partition comprised of a string of consecutive blocks which correspond to a certain segment of the partition, of which the allocation units is part of (and are therefore are smaller than partitions). In this case, partitions are strings of consecutive allocation units, each allocation unit is a string of consecutive blocks. In describing some embodiments of the present invention, for convenience purposes only, we relate to partitions as being the allocation units used in the cache, either explicitly or implicitly. However, it should be noted that the present invention is not limited in this respect; rather, some embodiments of the present invention may be implemented in conjunction with any cache using any kind of allocation units.
Reference is now made to
In accordance with some embodiments of the present invention, the cache controller 120 may include a groups management module 122 and a replacement procedure module 124. In accordance with further embodiments of the present invention, the cache controller 120 may include more than one replacement procedure module 124, as will be described below. Additional reference is now made to
In accordance with some embodiments of the present invention, the cache controller 120 may be adapted to utilize the replacement procedure module 124 to manage the cache in terms of groups of allocation units, rather than in terms of discrete allocation units (block 230). In accordance with further embodiments of the present invention, whenever it is necessary to discard or remove data from the cache 100, for example, when the cache's data space address 130 is full or is nearing full (the amount of data in the data space address crosses 130 a predetermined capacity threshold, for example) and storage space needs to be freed in order to make place for new data, the cache controller 120 may be adapted to utilize the replacement procedure module 124 as part of selecting which of the groups of allocation units is to be discarded. In accordance with yet further embodiments of the present invention, the cache controller 120 may be adapted to utilize the replacement procedure module 124 to create one or more queues, each of which associated with one or more groups of allocation units. In accordance with some embodiments of the present invention, the replacement procedure module 124 may associate each of the queues with groups of allocation units having a certain length or a certain range of lengths. The length of a group may be determined by the number of allocation units included in the group. The replacement procedure module 124 may be adapted to implement one or more of the following replacement procedures: LRU, FIFO, LFU, MFU or random selection, to create the one or more queues, as will be described in greater detail hereinbelow.
Reference is now made to
In
In accordance with some embodiments of the present invention, whenever a new allocation unit (e.g., AUn) is allocated, the cache controller 120, for example, may be configured to determine whether the allocation unit AUn-1 preceding the new allocation unit AUn is in the cache 100, and if in the cache 100, whether it was accessed while in the cache 100 (block 320). In accordance with some embodiments of the present invention, if it is determine that the allocation unit AUn-1 preceding the new allocation unit AUn is in the cache 100 and was accessed while in the cache 100, the new allocation unit AUn may be added to an existing group of allocation units Gm which includes at least allocation unit AUn-1, and possibly, one or more additional allocation units preceding AUn-1 (block 330). However, if it is determined that the allocation unit AUn-1 preceding the new allocation unit AUn is either not in the cache 100 or is in the cache 100 but was not accessed while in the cache 100, a new group of allocation units G1 may be created, and the new allocation unit AUn may be added the new group of allocation units G1 (block 335).
In accordance with some embodiments of the present invention, the groups management module 122, possibly in cooperation with the cache management module 120, may be adapted to determine whether a new allocation unit is associated with any of the existing groups of allocation units at least in accordance with a predefined criterion, and may be adapted to create a new group of allocation units when it is determined that a new allocation unit is not associated with any of the existing groups of allocation units. In accordance with further embodiments of the present invention, whenever a new group of allocation units is created, the groups management module 122 may be adapted to record data corresponding to the new group. In accordance with some embodiments of the present invention, the groups management module 122 may maintain a record for each group of allocation units specifying which allocation units belong that group. In accordance with further embodiments of the present invention, the groups management module 122 may be adapted to record in the metadata of each newly allocated allocation unit a reference to the group to which it belongs, such that all the allocation units in the cache 100 include such a reference.
In accordance with further embodiments of the present invention, the groups management module 122 may be further adapted to update the data records associated with any of the groups and/or with any (one or more) of the allocation units belonging to any of the groups whenever one of a set of predefined events which involve the group, or one (or more) of the allocation units belonging to the group, occurs. For example, as part of recording such an event, the time of occurrence of the event may be recorded in group's records and/or in the records of one or more of the allocation units belonging to the group. Examples of such events, may include, but are not limited to, the following: the addition of a new allocation unit to the group, the reading (accessing) of one of the allocation units belonging to the group while in the cache, etc. In accordance with further embodiments of the present invention, as part of recording or updating the records of a group of allocation units and/of one or more allocation units belonging to a certain group, the groups management module 122 may record a timestamp or a serial number (IDN), for example, corresponding to the time (absolute or relative) of occurrence of the event or any other predefined data associated with the event. It should be noted, that in accordance with some embodiments of the present invention, the groups management module 122 may be adapted to update the records of all of the allocation units belonging to a certain group, whenever a predefined event (or one of a set of predefined events) which involves one (or more) of the allocation units belonging to that group occurs. The groups management module 122 may also be configured to delete a group of allocation units and the data corresponding to the group when the groups is discarded from the cache. In accordance with some embodiments of the present invention, as part of discarding a group of allocation units, the groups management module 122, may be adapted to discard the entire group including all the allocation units belonging to the group, as will be described hereinbelow.
In accordance with some embodiments of the present invention, if the new allocation unit AUn has been added to an existing group of allocation units Gm, the group's group length value GmL, which represents the number of allocation units included in or belonging to the group, may be incremented (e.g., by the groups management module 122) by 1 (block 340), to accommodate for the addition of the new allocation unit and the corresponding growth in size. In accordance with some embodiments of the present invention, if, however, a new group of allocation units G1 has been created for the new allocation unit AUn, the new group's group length value GnL may be set to 1 (block 345), corresponding to the fact that there is only one allocation unit associated with or belonging to the new group.
In accordance with some embodiments of the present invention, the replacement procedure module 124 may be utilized (for example, by the cache management module 120) to create a plurality (two or more) of queues, for example LRU0, LRU1, LRU2, LRU3. In accordance with some embodiments of the present invention, the replacement procedure module 124 may be adapted to implement any replacement technique or procedure to create and manage any of the queues, including, but not limited to one or more of the following replacement procedures: LRU, FIFO, LFU, MFU or random selection. For convenience purposes, throughout the description of the present invention, reference is made to the LRU type replacement procedure, and in other places the use of an LRU is assumed, although not always specifically stated. However, it should be noted that the present invention is not limited to the use of any one particular replacement procedure and that any replacement procedure known in the present or yet to be devised in the future may be used as part of some embodiments of the present invention. Furthermore, it should be noted that some of the embodiments of the present invention described throughout the specification assume the use of an LRU replacement queue. It would be readily apparent to those of ordinary skill in the art to modify these embodiments of the present invention, if necessary, in cases where other replacement procedures are used, to maintain the spirit of the present invention.
In accordance with some embodiments of the present invention, each of the plurality of queues (LRU0, LRU1, LRU2, LRU3, for example) may be associated with groups of allocation units having a certain length or a certain range of lengths. For example, each of the queues may be associated with a particular group length value or with a particular range of group length values, such that the queue to which a certain group of allocation units is to be assigned may be determined in accordance with the length of the group of allocation units. In accordance with some embodiments of the present invention, once the group's group length value (GmL or GnL) has been updated (or set to 1), the groups management module 122 may be adapted to determine to with which of the plurality of queues, the group (Gm or Gn) to which the new allocation unit AUn belongs is to be assigned (block 350). In accordance with further embodiments of the present invention, the groups management module 122 may determine to which of the queues the group to which the new allocation unit belongs is to be assigned, in accordance with the group's group length value (GmL or GnL).
In accordance with some embodiments of the present invention, in case the new allocation unit AUn was added to an existing group of allocation units Gm, the group of allocation units Gm may either be left within the current queue to which it has been assigned previous to the addition of the new allocation unit or it may be moved to a different queue, as will be described below. For example, in case the group's incremented group length value (GmL) remains within the range of group length values with which the current queue is associated it will not be moved. However, if the group's incremented group length value (GmL) exceeds the maximum group length value with which the current queue is associated, then the group Gm may be transferred to a different queue. The queue to which the group Gm may be transferred may be the one which is associated with the incremented group length value (GmL).
In accordance with some embodiments of the present invention, once the replacement queue module 124, either independently or in cooperation with the groups management module 122, has determined to which of the queues (LRU0, LRU1, LRU2 or LRU3, for example) the group (Gm or Gn) of allocation units to which the new allocation unit AUn belongs is to be assigned, the replacement procedure module 124 may add the new allocation unit AUn to the tail of the queue which was determined to be the one to receive the group (Gm or Gn) of allocation units to which the new allocation unit AUn belongs (block 360). As part of adding the new allocation unit AUn to the selected queue, the allocation unit's metadata may be updated accordingly (for example, by the replacement procedure module 124). It should be noted that in the above embodiments of the present invention, the queues is assumed to implemented via an LRU replacement procedure, and accordingly, the new allocation unit is placed at the tail of the queue. However, it should also be noted, that the replacement queue module 124, may be otherwise configured to place the new allocation unit in a different position within the queue, for example, in case a different replacement procedure is implemented by the replacement queue module 124 to create and manage the replacement queue(s).
In accordance with some embodiments of the present invention, once the new allocation unit AUn is added to one of the queues, the replacement procedure 124, for example, in cooperation with the groups management module 122, may be adapted to determine whether all the allocation units belonging to the group (Gm or Gn) to which the new allocation unit AUn was added are already in the queue to which the new allocation units was added (block 370). As mentioned above, in accordance with some embodiments of the present invention, in case the new allocation unit AUn was added to an existing group of allocation units Gm, the group's incremented group length value (GmL) may either remain within the range of group length values of the queue with which the group Gm has been associated previous to the addition of the new allocation unit AUn, in which case, the group and/or the allocation units belonging to it may not be moved, or the group's incremented group length value (GmL) may exceed the maximum group length value of the queue with which the group Gm has been associated previous to the addition of the new allocation unit AUn, in which case the group and/or the allocation units belonging to it may be transferred to a different queue. The queue to which the group Gm may be transferred may be the one which is associated with the incremented group length value (GmL).
In accordance with some embodiments of the present invention, if it is determined that all the allocation units belonging to the group (Gm or Gn) to which the new allocation unit AUn was added are already in the queue to which the new allocation units was added, the replacement procedure module 124, for example, in cooperation with the groups management module 122, may be adapted to move all the allocation units belonging to the group (in this case only relevant for Gm) to the tail of the queue and, if necessary, update their metadata accordingly (block 380). However, in accordance with some embodiments of the present invention, if not all the allocation units belonging to the group (Gm or Gn) to which the new allocation unit AUn was added are already in the queue to which the new allocation units was added, the replacement procedure module 124, may be adapted to move all the allocation units belonging to the group (in this case only relevant for Gm) to the tail of the queue to which the new allocation unit was added, which will typically be the queue which is associated with the incremented group length value (GmL) (block 385).
As mentioned above, as part of the normal operation of the cache 100, at some point during the operation of the cache 100 (and usually, rather frequently) it may become necessary to discard data from the cache 100, for example, when the cache's data space address 130 is full or is nearing full. Reference is now made to
Reference is now additionally made to
As mentioned above, in accordance with some embodiments of the present invention whenever a new allocation unit (e.g. AUn) is allocated in the cache 100 a timestamp or a unique serial number (IDN) corresponding to the time (absolute or relative) of the allocation may be assigned to the new allocation unit. In accordance with further embodiments of the present invention, in case while in the cache, an allocation unit is accessed, the timestamp or the serial number associated with the accessed allocation unit may be replaced with a timestamp or a serial number corresponding to the time or instant in which the (absolute or relative) allocation unit was accessed. As part of assigning or replacing a timestamp or a unique serial number associated with an allocation unit, the timestamp or the unique serial number may be written into the allocation unit's metadata or the timestamp or the unique serial number in the allocation unit's metadata may be replaced, for example.
In accordance with some embodiments of the present invention, each of the allocation units in the cache 100 may be associated with a certain age or age value corresponding to the amount of time since the allocation unit was allocated or corresponding to the amount of time since the allocation units was most recently accessed. In accordance with some embodiments of the present invention, The age of an allocation unit at a certain point in time may be associated with the difference between the current clock or the current serial number (the one following the serial number assigned to the most recently allocation unit, for example) and the timestamp or the serial number in the allocation unit's metadata. For example, the age of an allocation unit may be determined by subtracting the serial number in the allocation unit's metadata from the current serial number.
In accordance with further embodiments of the present invention, an age value may be calculated for each group and not in-separate for each allocation unit in the cache. In accordance with some embodiments of the present invention, the age value of each of the groups may be determined in accordance with the age of one or more of the allocation units within that group. For example, in accordance with some embodiments of the present invention, the age of a certain group may be determined in accordance with, but not limited to, the following: the age of the youngest allocation unit in the group, the age of the oldest allocation unit in the group, the average age of all the allocation units in the group, etc. In accordance with some embodiments of the present invention, the age of the group of allocation units may be used to determine the age of one or more (e.g., all) the allocation units belonging to that group. For example, the age of the group of allocation units may be assigned to each of the allocation units belonging to that group. Throughout the specification and the claims, whenever reference is made to the age of an allocation units, unless explicitly or implicitly evident to the contrary, the age of the group of allocation units of which the allocation unit is part of part of, may be used instead.
Returning to
Once the ages of the heads of the queues are adjusted, the cache management module 120, for example, may be adapted to select from amongst the heads of the queues, the allocation unit (or the group of allocation units) having the highest adjusted age (block 430). The cache controller 120, may then determine, for example, in cooperation with the groups management module 122, to which group of allocation units the head of the queue which has been selected belongs (or is part of which) (block 440). Once the cache controller 120 determines to which group of allocation units the head of the queue which has been selected belongs, the cache controller may discard the entire group of allocation units, including all the allocation units belonging to that group (block 450).
Reference is now made to
One example of an operative criterion which may be utilized, according to some embodiments of the present invention, by the groups management module 122 to determine whether a new allocation unit is associated with an existing group of allocation units, and if so, with which of the existing groups of allocation units was already discussed above with reference to
It should be noted that as part of some embodiments of the present invention, any known in the present, or yet to be devised in the future, techniques, including various techniques which are well known in the present, may be used to implement the above operative criteria. As part of the implementation of one or more of the above operative criteria, the cache management module 120 and/or the groups management module 122, may be adapted to add various data to the metadata of one or more allocation units and/or to one or more data records associated with one or more of the groups of allocation units, including, but not limited to, various timestamps corresponding to various predefined events involving the allocation units or the groups of allocation units.
It should be noted that the above examples of various operative criteria used to determine whether an allocation unit should be added to an existing group of allocation units or not, are not exhaustive in nature, and that in accordance with further embodiments of the present invention the groups management module 122 may be configured to determine whether an allocation unit should be added to an existing group of allocation units in accordance with other operative criterion or criteria and combinations of various operative criteria, either in addition or in alternative to the operative criteria discussed above.
Those of ordinary skill in the art may appreciate that, in mass-storage systems, operatively related data (for example, data that is part of the same file) is often stored in consecutive locations in memory and/or is associated with consequtive serial numbers recognized throughout the stoage system. Due to this operative connection, and for other reasons, in mass-storage systems, consecutive data is often accessed together, for example, as part of processing a file which the consecutive data is part of. This concept is well known in the art and is sometimes referred to as “sequential locality”. Thus, by keeping consecutive, or otherwise related data together (either by maintaining the related data in the cache or by discarding it as a single unit) the efficiency of the cache may be improved.
In accordance with some embodiments of the present invention, in case, in accordance with at least the operative criterion, the groups management module 122 determines that the new allocation unit AUn is associated with one of the existing groups of allocation units (Gm), the groups management module 122 may be adapted to add the new allocation unit AUn to the existing group of allocation units Gm which was determined to be associated with the new allocation unit AUn (block 540). In accordance with further embodiments of the present invention, the groups management module 122 may be adapted to create a new group of allocation units (G1) in case it is determined that in accordance with the operative criterion, none of the existing groups of allocation units is appropriate to include the new allocation unit AUn. Once the groups management module 122 creates a new group of allocation units, the groups management module 122 may add (or otherwise associate) the new allocation unit AUn to the new group (block 545).
In accordance with some embodiments of the present invention, the groups management module 122 may include or may be associated with a storage device (not shown). In accordance with some embodiments of the present invention, the storage device may be used by the groups management module 122 to store a database including data relating to the groups of allocation units and to the allocation units belonging to each group. In accordance with some embodiments of the present invention, each entry in the database may be associated with one of the existing groups of allocation units, and may include, but not limited to including, data specifying which allocation units (one or more) belong to that group, the size of the group (the group's group length value, for example), and the aging factor associated with that group, which will be discussed in greater detail below.
In accordance with some embodiments of the present invention, if the new allocation unit AUn has been added to an existing group of allocation units Gm, the group's group length value GmL, which represents the number of allocation units included in or belonging to the group, may be incremented (e.g., by the groups management module 122) by 1 (block 550), to accommodate for the addition of the new allocation unit AUn and the corresponding growth in size. In accordance with some embodiments of the present invention, if, however, a new group of allocation units G1 has been created for the new allocation unit AUn, the new group's group length value GnL may be set to 1 (block 555), corresponding to the fact that there is only one allocation unit associated with or belonging to the new group.
In accordance with some embodiments of the present invention, as part of adding the new allocation unit to a group of allocation units (whether an existing group Gm or a newly created group G1), in addition to updating the group length value, the unique serial number (IDN) associated with the new allocation unit AUn may be assigned to the group of allocation units (Gm or G1) to which the new allocation unit AUn was added and/or to each of the allocation units belonging to the group of allocation units to which the new allocation units was added (this option is application only to the existing group of allocation units Gm) (block 560). In accordance with some embodiments of the present invention, the unique serial number, IDN, is the serial number assigned to the new allocation unit upon the allocation thereof. However, as mentioned above, in accordance with further embodiments of the present invention, instead of the (relative) unique serial number (IDN), a timestamp, for example, corresponding to the time (absolute or relative) of occurrence of the allocation of the new allocation unit may be assigned to the new allocation unit, in which case, the timestamp associated with the new allocation unit AUn may be assigned to the group of allocation units (Gm or G1) to which the new allocation unit AUn was added and/or to each of the allocation units belonging to the group of allocation units to which the new allocation units was added.
In accordance with some embodiments of the present invention, the assignment of a unique serial number (IDN) or of a timestamp to a certain allocation unit, and as a result, to the group of allocation units to which that allocation units belongs and/or to each of the allocation units belonging to the group of allocation units to which the that allocation units belongs may occur in response to other events, in addition to the allocation of the allocation unit. For example, a new IDN or a timestamp may be assigned to an allocation unit whenever the allocation unit is accessed while in the cache. Thus, in accordance with some embodiments of the present invention, a new IDN or a timestamp corresponding to the (time of occurrence of) accessing of a certain allocation unit while in the cache may be assigned to the group of allocation units to which that allocation units belongs and/or to each of the allocation units belonging to the group of allocation units to which the that allocation units belongs.
Since, as was discussed above and as will be further described below, in accordance with some embodiments of the present invention, the discard priority of the groups of allocation units (or of the groups of allocation units which are the candidates for being discarded) may be determined in accordance with (amongst other factors) the serial number of the timestamp associated with each of the groups and/or with the serial number or the timestamp associated with one or more (e.g., each of) the allocation units belonging to each of the groups, a recent accessing of an allocation unit belonging to a certain allocation unit may change the group's priority to be discarded. For example, in accordance with some embodiment of the present invention, in case an LRU replacement procedure is implemented in the cache 100, and a certain group is assigned with a serial number of a timestamp corresponding to a recent event involving one of the allocation units belonging to that group, that group may receive a lower priority to be discarded due to its association with the serial number or timestamp corresponding to the substantially recent event.
Reference is now made to
In accordance with some embodiments of the present invention, the cache management module 120, may be adapted to adjust the age of each of the groups of allocation units in accordance with the aging factors associated with each of the group's group length value (unless for one or more of the groups, the aging factor is 1, in which case, the age will remain unchanged) (block 575). In accordance with some embodiments of the present invention, if in accordance with the policy implemented in the cache 100, the larger a group's group length value(s) is, the larger the aging factor which may be associated with that group, the age of the longer or larger groups may be modified to a greater extent since these groups are associated with the larger aging factors. However, as stated above, the present invention is not limited to this policy.
Once, the ages of the groups are adjusted, the cache management module 120, for example, may be adapted to select from amongst all the groups of allocation units, the group of allocation units having the highest adjusted age (block 580). The cache management module 120, may then determine, for example, in cooperation with the groups management module 122, which allocation units belong to or are included in that group, and may discard all the allocation units belonging to that group (block 585).
While certain features of the invention have been illustrated and described herein, many modifications, substitutions, changes, and equivalents will now occur to those skilled in the art. It is, therefore, to be understood that the appended claims are intended to cover all such modifications and changes as fall within the true spirit of the invention.
This application claims priority from U.S. Provisional Patent Application No. 60/639,780, filed Dec. 29, 2004, which is hereby incorporated by reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| 60639780 | Dec 2004 | US |
