WO1997008621A1 - System and method for sequential detection in a cache management system - Google Patents
System and method for sequential detection in a cache management system Download PDFInfo
- Publication number
- WO1997008621A1 WO1997008621A1 PCT/US1996/013951 US9613951W WO9708621A1 WO 1997008621 A1 WO1997008621 A1 WO 1997008621A1 US 9613951 W US9613951 W US 9613951W WO 9708621 A1 WO9708621 A1 WO 9708621A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- record
- track
- information field
- current track
- state
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F12/00—Accessing, addressing or allocating within memory systems or architectures
- G06F12/02—Addressing or allocation; Relocation
- G06F12/08—Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
- G06F12/0802—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches
- G06F12/0866—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches for peripheral storage systems, e.g. disk cache
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F12/00—Accessing, addressing or allocating within memory systems or architectures
- G06F12/02—Addressing or allocation; Relocation
- G06F12/08—Addressing or allocation; Relocation in hierarchically structured memory systems, e.g. virtual memory systems
- G06F12/0802—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches
- G06F12/0862—Addressing of a memory level in which the access to the desired data or data block requires associative addressing means, e.g. caches with prefetch
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2212/00—Indexing scheme relating to accessing, addressing or allocation within memory systems or architectures
- G06F2212/31—Providing disk cache in a specific location of a storage system
- G06F2212/312—In storage controller
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2212/00—Indexing scheme relating to accessing, addressing or allocation within memory systems or architectures
- G06F2212/60—Details of cache memory
- G06F2212/6026—Prefetching based on access pattern detection, e.g. stride based prefetch
Definitions
- This invention relates generally to memory subsystems for the storage of data in data processing systems. More particularly, the invention relates to an improved system and method for detecting sequentiality in memory accesses.
- Data caching is a well known technique for eliminating delays in memory access due to mechanical limitations of a storage device.
- a disk drive plural disks rotate at a fixed speed past read/write heads which may either be stationary with respect to the disk or move radially back and forth with respect to the disk in order to juxtapose the heads to various portions of the disk surfaces.
- This "access" time includes the time for a head to move to the correct cylinder (seek time) and the time required (or latency) for the disk to rotate with respect to the head until the beginning of the particular record sought is juxtaposed to the head for reading and writing.
- Cache data storage eliminates these inherent delays by storing records in frequently accessed tracks in a high speed system memory (e.g., solid-state RAM).
- a high speed system memory e.g., solid-state RAM
- the idea is simply to allow as many memory accesses as possible to immediately retrieve data from the high speed system memory rather than wait for the data to be transferred (or staged) from the slower disk storage device to the high speed system memory.
- data may be staged into the high speed system memory before data access is required (i.e., prestaged).
- the effectiveness of the cache data storage system is limited by the system's ability to anticipate the needs of future memory accesses and transfer those data records from disk storage to the high speed system memory prior to the memory access.
- These general considerations are described in Houtemaker et al., "MVS I/O Subsystems: Configuration Management and Performance Analysis", McGraw-Hill, 1993, which is hereby incorporated by reference in its entirety.
- the cache data storage system cannot anticipate future memory accesses.
- one method of anticipating future memory accesses is to identify sequential or near sequential memory accesses. Once a sequential or near sequential access is identified, future records/tracks in the sequence can be immediately prestaged into the high speed system memory in advance of future memory accesses.
- the system utilizes "hints" provided by the host operating system to go into sequential mode.
- hints provided by the host operating system to go into sequential mode. Examples of disk storage systems which rely on hints from the host operating system include the Storage Technology Corporation Iceberg RAID data storage subsystem and the International
- the limited size of cache memory directly influences the number of tracks that may be speculatively staged into the high speed system memory. This follows since, generally, the cache memory is a costly resource that must be managed properly to increase system performance. As a further consequence, the limited size of cache memory increases the number of acceptable "misses" prior to finding sequentiality. In this context, a "miss" is an access to a record that has not been prestaged into the high speed system memory. Clearly, prestaging of tracks should not occur unless there exists a sufficient probability that those staged tracks will be the subject of future memory accesses. Finally, the overhead required to administer the sequential detection system must not become overly burdensome wherein the performance of the cache data storage system itself is compromised.
- the invention satisfies the above mentioned needs by providing a low overhead system that detects sequentiality with a minimal number of cache "misses".
- the invention provides a system and method for tracking memory accesses at the record level as opposed to the track level. By analyzing the memory accesses at a finer granularity without excessively increasing the overhead, the system minimizes the number of "misses" prior to detecting sequentiality.
- the invention first maintains within system memory, a cache directory entry for each track.
- the directory entry includes a data structure comprising four data fields.
- the four data fields include the last sector/record referenced (LSRR), the state of the track (ST), the number of pre-stages already requested (PAR), and the stage ahead limit (SAL).
- LSRR last sector/record referenced
- ST state of the track
- PAR pre-stages already requested
- SAL stage ahead limit
- the system and method for sequential detection first initializes the ST field of a track to an initial state prior to being accessed for the first time.
- the ST field of the track is set to a first nonsequential state and the LSRR field is set to the number of the record that was accessed.
- the LSRR field is set to the value n.
- the ST field of the track is set to a second nonsequential state and the LSRR field is set to the value (n+l).
- the ST field of the track is set to a sequential state.
- d e LSRR field may alternatively store sector numbers instead of record numbers.
- the LSRR field would store the sector number of the end of the last transfer done on the track in the previous access.
- a sequential state may be reached after accessing two consecutive records only.
- a track may immediately enter a sequential state based on hints provided by the host system. Specifically, a sequential state is reached if either (1) the sequential bit is set in a Define Extent command, (2) a Read Track command is received, (3) a Read Multiple Count Key Data command is received, or (4) a Write Count Key and Data Next Track command is received.
- Figure 1 shows a high level diagram of a disk array system that incorporates the sequential detection system and method of the present invention
- Figure 2 shows a state diagram that identifies the process of identifying sequentiality in memory access requests
- Figures 3 and 4 show the effect of adjustment parameters on staging and destaging of tracks after sequentiality has been detected.
- FIG. 1 shows a simplified illustration of a cache management system 100.
- Cache management system 100 includes host directors 102a- 102b that are adapted to receive data read/write commands and associated data over one or more communication channels 103a, 103b from one or more hosts 110a- 110b.
- Host directors 102a- 102b are coupled to system memory 108 via memory controller 106.
- Host directors 102a-102b are coupled to memory controller 106 by means of bus 120.
- Cache management system 100 also includes device controllers 104a- 104b.
- Device controllers 104a- 104b are coupled to one or more storage devices 112a- 112b and serve to control the reading and writing of data to or from the respective data storage devices 112a- 112b.
- Device controllers 104a- 104b are also coupled to memory controller 106 by means of bus 120.
- cache management system 100 reduces access time by a host 110a- 110b to data stored in storage devices 112a- 112b by storing data temporarily in system memory 108.
- a host 110a- 110b can then access data from system memory 108 without suffering the inherent delays associated with the seek and latency times of mechanical storage devices 112a-112b.
- cache management system 100 must accurately predict what data to prestage into system memory 108.
- the present invention provides an improved method of accurately identifying these sequential memory accesses while suffering a minimal number of "misses" .
- the system and method for sequential detection relies on either internal memory access analysis or "hints" from host systems 110a- 110b.
- the system and method can be divided into the following functional parts: detection of sequentiality, operation in sequential mode, and termination of sequential mode. Each of these parts will be discussed separately in detail below.
- the detection of sequentiality is improved over known systems by analyzing a sequence of memory accesses at the record level rather than the track level. This approach minimizes the number of "misses" by identifying sequentiality within a single track rather than across multiple tracks.
- sequentiality is detected only after a sequence of multiple tracks are accessed, the first record accessed in each of the initial tracks is accessed without the benefit of being prestaged in the cache data storage system.
- merely refining the sequential detection to a smaller scale does not automatically produce a more efficient cache data storage system.
- the increased overhead required for analysis at a finer granularity degrades performance since one can reasonably expect an order of magnitude increase in the overhead costs involved in the monitoring and analysis of memory accesses.
- any method that identifies sequentiality faster than known systems must insure that a plurality of tracks are not staged prematurely.
- the present invention balances these concerns by providing a system and method for analyzing memory accesses at the record level without substantial increases in overhead. As a consequence, the sequential detection system of the present invention minimizes the number of misses prior to the detection of sequentiality without significantly impacting system performance.
- the system and method for sequential detection stores, in system memory, a data structure in a directory entry for each track.
- the data structure is divided into four 8-bit fields that include the Last Sector/Record Referenced (LSRR), State (ST), Pre-stages Already Requested (PAR), and Stage Ahead Limit (SAL) information fields.
- LSRR and ST fields are used primarily in the sequential detection phase.
- the PAR and the SAL fields are utilized after sequentiality has been detected.
- the LSRR and ST fields are described in detail in this section below. An operative description of the PAR and SAL fields is given in a following section.
- the LSRR information field contains the last record number referenced on the track.
- the LSRR information field contains the cache sector number of the end of the last transfer done on the track.
- a sector is a fixed length portion of memory on the disk. The specific choice between these two options depends on the actual system implementation. For example, if record numbers are used, a non-standard track layout such as non- consecutively ascending record numbers may not be seen as sequential. On the other hand, if sector numbers are used, the sequential test will have to consider key fields, skipped sectors, etc. to identify an acceptable range between accessed sectors that can be recognized as part of a sequence of sequential accesses.
- the ST information field contains the state of the corresponding track. The possible values of the ST field are listed in Table 1. A more detailed description of these states as they relate to Figure 2 is given in more detail below.
- the PAR information field identifies the number of tracks that have already been requested ahead of this track. This field is set when a stage ahead request is made after detection of sequentiality. Finally, the SAL information field identifies the number of tracks that the system desires to be staged ahead relative to that track.
- the internal sequential detection system begins by setting the ST field of a track equal to 0 when a device controller 104a- 104b completes the staging of that track. This step is illustrated by transition 202 in Figure 2.
- host director 102a- 102b subsequently accesses a staged track and finds ST equal to 0, host director 102a- 102b checks the directory entry for the previous track (unless the current track is Cylinder 0 Head 0). If the ST field in the previous track did not have the high order bit set (i.e. , ST ⁇ 128) then host director 102a- 102b changes the ST field in the current track from state 0 to state 1 and sets the LSRR field to the number of the record that was accessed.
- the LSRR field is set to the sector number of the end of the last transfer done on the track. This step is illustrated by transition 204.
- the first access to a particular track merely serves notice that a memory access to a record on the track has occurred. At this point, even if the first memory access to the track is the start of a sequence, the memory access is indistinguishable from a random memory access.
- the previous track was in cache and the ST information field did have the high order bit set (i.e., ST > 128), then sequentiality is detected and the ST field for the current track is changed from state 0 to state 128 (transition 212). This scenario may occur in the following situations.
- the record accessed in the current track could be the first positioned record in the track that follows after the last positioned record in the previous track. In other words, the access has continued sequentially from the previous track to the current track.
- the record accessed in the current track does not directly follow the last record accessed in the previous track (i.e. , near sequential accessing). In either case, sequentiality is detected after the first access into the current track.
- the internal sequential detection system must await future accesses to that particular track before detecting sequentiality. Specifically, upon a subsequent access to the track, host director 102a- 102b checks the LSRR field in the directory entry. If the record number sought to be accessed is one greater than the record number stored in the LSRR field (or the sector number is within an acceptable range of the last sector accessed as indicated by the LSRR field), then the ST field is changed from state 1 to state 2 (transition 206) and the LSRR field is updated to the appropriate sector/record number. State 2 represents an intermediate state within the sequentiality hierarchy.
- additional intermediate stages may be used to further minimize false positives. Whether such additional stages are necessary depends upon the size, position, etc. of records within the track in light of the general tradeoff between prestaging too many tracks too early and incurring "misses" by waiting too long.
- transition 210 between state 1 and state 128 represents an exception to the necessity of utilizing intermediate state 2.
- This exception is based purely on the size of the records being accessed. For example, if the record sizes are larger than 15,476 bytes (i.e., maximum record size for 3 records/ track in the International Business Machines 3380 disk family) then a subsequent access to a track will prompt the ST field to be changed from state 1 to state 128 (i.e., sequentiality detected).
- the exact threshold of the record size (or maximum records/track) may vary depending upon a particular system implementation and application.
- the sequential detection method requires a second confirmation before sequentiality is indicated/detected.
- the ST field is changed from state 2 to state 128 (transition 208), indicating the detection of sequentiality.
- states 129-132 describe an aspect of the sequential detection system that relies on hints from host system HOa-llOb. Transitions 222, 224, 226, 228 to any of states 129-132 may occur at any time during the flow within the internal sequential detection system. Thus, the ST field of a track may transition from any of states 0-2 to any of states 129-132.
- transition 222 to state 129 occurs when the sequential bit is set in the Define Extent command
- transition 224 to state 130 occurs when a Read Track command is received
- transition to state 131 occurs when a
- State 132 occurs when a Write Count Key and Data Next Track command is received.
- states 129-132 the high order bit is set and sequentiality is detected.
- a similar transition would occur based on the contents of the parameters of a Locate Record command. For example, by using the count and transfer length factor of the Locate Record command, the sequential detect system can identify how many tracks should be prestaged. Similarly, in an alternative embodiment, the sequential detect system utilizes the count in the Locate Record command and the block size in the Define
- cache management system copies the contents from cache to disk.
- deallocating cache management system makes an area of cache available for future use. Generally, if a write has occurred, a track is destaged before being deallocated. If only reads have occurred, a track may simply be deallocated.
- a stage ahead decision is made at the end of a locate record domain and/or the command chain.
- five dynamic adjustment parameters (K1-K5) are defined as follows:
- the KI, K2 and K5 adjustment parameters control the staging process.
- the values of these adjustment parameters depend on such factors as the size of the records (tracks with smaller records require more time to process since more records exist on a track), the speed of the channels, etc. that influence the speed of track processing. Simply stated, the faster the speed of track processing, the greater the number of tracks that should be prestaged into system memory 108. However, because the internal sequential detection system has no way to predict how far the sequential access will continue, there is the continuing risk of staging ahead too far.
- the K5 adjustment parameter directly addresses this issue.
- the value of this adjustment parameter is similar to KI (subsequent requests) but is used the first time sequentiality is detected.
- the initial value of KI is eight and the initial value of K5 is three.
- the stage ahead request based on a first detection of sequentiality, would cause three tracks rather than eight tracks to be prestaged into system memory 108.
- An example of this situation is transition 208 between states 2 to 128 in Figure 2. In this transition, a first sequential detection is identified and three tracks are prestaged into system memory 108.
- the value of K5 is used to address the case of short runs that may not require the full length of a subsequent prestage request of eight tracks.
- Figure 3 shows a sequence of staging/destaging that relies on adjustment parameters K1-K4 described above.
- ST 129
- PAR fields for tracks 145-150 show a decreasing sequence of numbers that represent the number of tracks staged ahead relative to that particular track.
- the PAR value of three is still above the K2 value of two. Accordingly, no subsequent stage ahead request is sent.
- the PAR value equals K2.
- the SAL field is referenced to identify the number of tracks to stage ahead. Since the SAL field contains the KI value, this example would require that eight more tracks be staged.
- the PAR value of track 151 is updated to ten.
- adjustment parameters KI, K2 and K5 identify when and how far to stage ahead. To facilitate dynamic adjustment, the number of prestaged tracks that have not been accessed is monitored. If the number of non-accessed tracks becomes excessive, then adjustment parameters KI, K2 or K5 can be selectively decreased.
- the specific method and thresholds for dynamic adjustment are implementation specific.
- Adjustment parameter K3 identifies the track behind the current track that can be deallocated, if the current track hasn't been read.
- the initial value of K3 is three.
- Adjustment parameter K4 is analogous to adjustment parameter K3 but is used if the current track or the one to be destaged has been written to.
- the initial value of adjustment parameter K4 is sixteen (or one cylinder) . This particular value is implementation dependent. This value was chosen in order to prevent the data drive for destages from being the data drive that is being staged. By this design, reading and writing occur on separate physical disks.
- adjustment parameter K4 is dynamically adjusted to prevent a parity drive on destages from interfering with the stages.
- Figure 4 illustrates the operation of adjustment parameters K3 and K4. As illustrated, if the current track is track 145, a K3 value of three would cause track 142 to move to the least recently used (LRU) memory if current track 145 has not been read. Similarly, a K4 value of 16 (i.e. , one cylinder) would indicate that track 140 would be destaged behind. Termination of Sequential Mode
- cache management system 100 efficiently identifies when to terminate sequential mode.
- the slots in system memory 108 that contain staged tracks are logically linked based on the time of their most recent usage.
- the slots at the end of the logical chain define the least recently used (LRU) slots.
- the K3 adjustment parameter effectively controls termination.
- the system is designed to anticipate when sequential accessing has stopped in order to get the unread cache slots available sooner.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Memory System Of A Hierarchy Structure (AREA)
Abstract
Description
Claims
Priority Applications (4)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
JP9510587A JPH11506238A (en) | 1995-08-29 | 1996-08-29 | System and method for sequentiality detection in a cache management system |
AU69610/96A AU707876B2 (en) | 1995-08-29 | 1996-08-29 | System and method for sequential detection in a cache management system |
DE69609515T DE69609515T2 (en) | 1995-08-29 | 1996-08-29 | METHOD FOR SEQUENTIAL DETECTION IN A CACHE STORAGE MANAGEMENT SYSTEM |
EP96930636A EP0847555B1 (en) | 1995-08-29 | 1996-08-29 | Method for sequential detection in a cache management system |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US08/521,026 US5867685A (en) | 1995-08-29 | 1995-08-29 | System and method for sequential detection in a cache management system |
US08/521,026 | 1995-08-29 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO1997008621A1 true WO1997008621A1 (en) | 1997-03-06 |
Family
ID=24075023
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/US1996/013951 WO1997008621A1 (en) | 1995-08-29 | 1996-08-29 | System and method for sequential detection in a cache management system |
Country Status (7)
Country | Link |
---|---|
US (1) | US5867685A (en) |
EP (1) | EP0847555B1 (en) |
JP (1) | JPH11506238A (en) |
AU (1) | AU707876B2 (en) |
CA (1) | CA2230542A1 (en) |
DE (1) | DE69609515T2 (en) |
WO (1) | WO1997008621A1 (en) |
Families Citing this family (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6381677B1 (en) * | 1998-08-19 | 2002-04-30 | International Business Machines Corporation | Method and system for staging data into cache |
US6449697B1 (en) * | 1999-04-23 | 2002-09-10 | International Business Machines Corporation | Prestaging data into cache in preparation for data transfer operations |
US6622212B1 (en) * | 1999-05-24 | 2003-09-16 | Intel Corp. | Adaptive prefetch of I/O data blocks |
US6865514B1 (en) * | 2000-08-18 | 2005-03-08 | Emc Corporation | Post test data processing of a mass storage system |
AUPR611101A0 (en) * | 2001-07-04 | 2001-07-26 | Commonwealth Scientific And Industrial Research Organisation | Apparatus and analysis method for determination of moisture in materials |
US7177853B1 (en) * | 2002-02-21 | 2007-02-13 | Emc Corporation | Cache management via statistically adjusted time stamp queue |
US6948009B2 (en) | 2002-06-04 | 2005-09-20 | International Business Machines Corporation | Method, system, and article of manufacture for increasing processor utilization |
US7353334B2 (en) * | 2002-08-19 | 2008-04-01 | Aristos Logic Corporation | Method of increasing performance and manageability of network storage systems using optimized cache setting and handling policies |
US7124128B2 (en) * | 2003-06-17 | 2006-10-17 | International Business Machines Corporation | Method, system, and program for managing requests to tracks subject to a relationship |
US7099995B2 (en) * | 2004-02-17 | 2006-08-29 | International Business Machines Corporation | Metadata access during error handling routines |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0109309A2 (en) * | 1982-11-15 | 1984-05-23 | Storage Technology Corporation | Method of operating a computer system |
US5146578A (en) * | 1989-05-01 | 1992-09-08 | Zenith Data Systems Corporation | Method of varying the amount of data prefetched to a cache memory in dependence on the history of data requests |
JPH0535573A (en) * | 1991-03-25 | 1993-02-12 | Nec Corp | Remote file access system |
Family Cites Families (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US4875155A (en) * | 1985-06-28 | 1989-10-17 | International Business Machines Corporation | Peripheral subsystem having read/write cache with record access |
US5253351A (en) * | 1988-08-11 | 1993-10-12 | Hitachi, Ltd. | Memory controller with a cache memory and control method of cache memory including steps of determining memory access threshold values |
US5201053A (en) * | 1990-08-31 | 1993-04-06 | International Business Machines Corporation | Dynamic polling of devices for nonsynchronous channel connection |
US5381539A (en) * | 1992-06-04 | 1995-01-10 | Emc Corporation | System and method for dynamically controlling cache management |
US5426761A (en) * | 1992-12-22 | 1995-06-20 | International Business Machines Corporation | Cache DASD sequential staging and method |
-
1995
- 1995-08-29 US US08/521,026 patent/US5867685A/en not_active Expired - Lifetime
-
1996
- 1996-08-29 JP JP9510587A patent/JPH11506238A/en active Pending
- 1996-08-29 CA CA002230542A patent/CA2230542A1/en not_active Abandoned
- 1996-08-29 AU AU69610/96A patent/AU707876B2/en not_active Ceased
- 1996-08-29 DE DE69609515T patent/DE69609515T2/en not_active Expired - Fee Related
- 1996-08-29 WO PCT/US1996/013951 patent/WO1997008621A1/en active IP Right Grant
- 1996-08-29 EP EP96930636A patent/EP0847555B1/en not_active Expired - Lifetime
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
EP0109309A2 (en) * | 1982-11-15 | 1984-05-23 | Storage Technology Corporation | Method of operating a computer system |
US5146578A (en) * | 1989-05-01 | 1992-09-08 | Zenith Data Systems Corporation | Method of varying the amount of data prefetched to a cache memory in dependence on the history of data requests |
JPH0535573A (en) * | 1991-03-25 | 1993-02-12 | Nec Corp | Remote file access system |
Non-Patent Citations (2)
Title |
---|
ANONYMOUS: "Improving Performance of Buffered DASD to which some References are Sequential. August 1981.", IBM TECHNICAL DISCLOSURE BULLETIN, vol. 24, no. 3, August 1981 (1981-08-01), NEW YORK, US, pages 1559 - 1562, XP002020746 * |
PATENT ABSTRACTS OF JAPAN vol. 017, no. 326 (P - 1560) 21 June 1993 (1993-06-21) * |
Also Published As
Publication number | Publication date |
---|---|
JPH11506238A (en) | 1999-06-02 |
DE69609515T2 (en) | 2000-12-14 |
EP0847555B1 (en) | 2000-07-26 |
MX9801539A (en) | 1998-10-31 |
DE69609515D1 (en) | 2000-08-31 |
EP0847555A1 (en) | 1998-06-17 |
AU707876B2 (en) | 1999-07-22 |
AU6961096A (en) | 1997-03-19 |
US5867685A (en) | 1999-02-02 |
CA2230542A1 (en) | 1997-03-06 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
US4875155A (en) | Peripheral subsystem having read/write cache with record access | |
EP0077451B1 (en) | Storage subsystem including a bypassable cache | |
US5809560A (en) | Adaptive read-ahead disk cache | |
US6311252B1 (en) | Method and apparatus for moving data between storage levels of a hierarchically arranged data storage system | |
EP0301211B1 (en) | Cache management for a peripheral data storage subsystem | |
EP0077453B1 (en) | Storage subsystems with arrangements for limiting data occupancy in caches thereof | |
US5253351A (en) | Memory controller with a cache memory and control method of cache memory including steps of determining memory access threshold values | |
US5134563A (en) | Sequentially processing data in a cached data storage system | |
EP0077452B1 (en) | Data promotion in storage subsystems | |
EP0109309B1 (en) | Method of operating a computer system | |
US6880043B1 (en) | Range-based cache control system and method | |
AU673488B2 (en) | Cache memory system and method of operating the cache memory system | |
US4956803A (en) | Sequentially processing data in a cached data storage system | |
US6341331B1 (en) | Method and system for managing a raid storage system with cache | |
US5991775A (en) | Method and system for dynamic cache allocation between record and track entries | |
US7120759B2 (en) | Storage system and method for prestaging data in a cache for improved performance | |
US7437515B1 (en) | Data structure for write pending | |
US5694570A (en) | Method and system of buffering data written to direct access storage devices in data processing systems | |
US5867685A (en) | System and method for sequential detection in a cache management system | |
US6336164B1 (en) | Method and system for preventing deadlock in a log structured array | |
US7139874B1 (en) | System and method for providing client-directed staging to improve non-sequential access performance in a caching disk storage system | |
JPH08263380A (en) | Disk cache control system | |
JP4104283B2 (en) | Storage subsystem and information processing system | |
JPH07134634A (en) | Disk device | |
JPH07334425A (en) | Disk cache control system |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
AK | Designated states |
Kind code of ref document: A1 Designated state(s): AU CA JP MX |
|
AL | Designated countries for regional patents |
Kind code of ref document: A1 Designated state(s): AT BE CH DE DK ES FI FR GB GR IE IT LU MC NL PT SE |
|
DFPE | Request for preliminary examination filed prior to expiration of 19th month from priority date (pct application filed before 20040101) | ||
121 | Ep: the epo has been informed by wipo that ep was designated in this application | ||
ENP | Entry into the national phase |
Ref document number: 2230542 Country of ref document: CA Ref country code: CA Ref document number: 2230542 Kind code of ref document: A Format of ref document f/p: F |
|
WWE | Wipo information: entry into national phase |
Ref document number: PA/a/1998/001539 Country of ref document: MX |
|
ENP | Entry into the national phase |
Ref country code: JP Ref document number: 1997 510587 Kind code of ref document: A Format of ref document f/p: F |
|
WWE | Wipo information: entry into national phase |
Ref document number: 1996930636 Country of ref document: EP |
|
WWP | Wipo information: published in national office |
Ref document number: 1996930636 Country of ref document: EP |
|
WWG | Wipo information: grant in national office |
Ref document number: 1996930636 Country of ref document: EP |