公开(公告)号：US5717888A

公开(公告)日：1998-02-10

申请号：US459864

申请日：1995-06-02

申请人： Susan Kay Candelaria ,  Joseph Smith Hyde ,  Vernon John Legvold

发明人： Susan Kay Candelaria ,  Joseph Smith Hyde ,  Vernon John Legvold

IPC分类号： G06F12/08 ,  G06F12/12

CPC分类号： G06F12/0866

摘要： In a data storage system having a direct access storage device (DASD) and a cache, a cache directory has two types of directory entries. A track directory entry TDE identifies up to one DASD track of data records currently stored in cache. All records stored in a DASD track can be stored in a cache storage space allocated for the DASD track identified by the TDE. A cylinder directory entry CDE identifies a number N of records from any track in a respective cylinder of tracks. N is a positive integer less than the total number of records storable in a DASD track. From one to all of the DASD tracks in one cylinder may be identified in a CDE. The cache data storage allocation corresponding to a CDE is the same as that allocated for a TDE. Each TDE is addressed by a DASD track address of a cylinder while a CDE is addressed using a pseudo track number corresponding to a servo track in the DASD. A record cast out control for the cache includes examining the number of records destaged. Based on the number of records cast out from cache to DASD, a cache event parameter CEP is modified or maintained. Accessing cache requires examining a current CEP value for each accessed DASD track to access either the TDE or CDE.

摘要翻译： 在具有直接访问存储设备（DASD）和高速缓存的数据存储系统中，高速缓存目录具有两种类型的目录条目。 轨道目录条目TDE最多识别当前存储在高速缓存中的数据记录的一个DASD轨道。 存储在DASD轨道中的所有记录可以存储在由TDE标识的DASD轨道分配的高速缓存存储空间中。 圆柱体目录条目CDE从轨道的相应圆柱体中的任何轨道识别N个记录。 N是小于在DASD轨道中可存储的记录总数的正整数。 在一个气缸中的一个到所有DASD轨迹可以在CDE中被识别。 与CDE对应的缓存数据存储分配与分配给TDE的缓存数据存储分配相同。 每个TDE由气缸的DASD轨道地址寻址，而使用对应于DASD中的伺服轨道的伪轨道号来寻址CDE。 高速缓存的记录丢弃控制包括检查已删除的记录数。 根据从缓存转发到DASD的记录数，缓存事件参数CEP被修改或维护。 访问缓存需要检查每个访问的DASD磁道的当前CEP值以访问TDE或CDE。

公开(公告)号：US5815656A

公开(公告)日：1998-09-29

申请号：US949668

申请日：1992-09-23

申请人： Susan Kay Candelaria ,  Michael Howard Hartung ,  Dennis Albert Kukula ,  Kenneth Wayne Lane ,  Vernon John Legvold ,  Guy Eugene Martin ,  John Norbert McCauley, Jr. ,  Carol Santich Michod ,  Mark Albert Reid ,  William Lee Richardson

发明人： Susan Kay Candelaria ,  Michael Howard Hartung ,  Dennis Albert Kukula ,  Kenneth Wayne Lane ,  Vernon John Legvold ,  Guy Eugene Martin ,  John Norbert McCauley, Jr. ,  Carol Santich Michod ,  Mark Albert Reid ,  William Lee Richardson

IPC分类号： G06F3/06 ,  G06F11/14 ,  G06F12/08 ,  G06F11/00

CPC分类号： G06F11/0793 ,  G06F11/0727 ,  G06F11/201 ,  G06F12/0866

摘要： A data processing system executes a plurality of processes in parallel. The processes access shared user records stored in memory serially. One or more data structures are used to track access by processes to the data structures and to the user records. Responsive to a fault occurring in a first parallel process, a signal is given the remaining parallel processes indicating the failure. Responsive to the failure, it is determined if the faulting parallel process had access to one of the data structures. Depending upon the circumstances of access, integrity of the user records may be assumed and the record validated. The data structures themselves can be examined to determine if the contents of memory is reliable. Typically the data structures support use of the memory as a cache.

摘要翻译： 数据处理系统并行地执行多个处理。 这些进程以串行方式访问存储在存储器中的共享用户记录。 一个或多个数据结构用于跟踪对数据结构和用户记录的进程的访问。 响应于在第一并行过程中发生的故障，给出指示故障的剩余并行过程的信号。 响应于失败，确定故障并行过程是否可以访问其中一个数据结构。 根据访问的情况，可以假设用户记录的完整性并且记录被验证。 可以检查数据结构本身以确定存储器的内容是否可靠。 通常，数据结构支持使用内存作为缓存。

公开(公告)号：US5682513A

公开(公告)日：1997-10-28

申请号：US414826

申请日：1995-03-31

申请人： Susan Kay Candelaria ,  Vernon John Legvold ,  Warren Keith Stanley

发明人： Susan Kay Candelaria ,  Vernon John Legvold ,  Warren Keith Stanley

IPC分类号： G06F11/14 ,  G06F12/08 ,  G06F12/02 ,  G06F12/16

CPC分类号： G06F11/2064 ,  G06F11/2074 ,  G06F11/2076 ,  G06F12/0866

摘要： A data storage system provides disaster recovery capability by asynchronously transmitting record updates to a secondary site in sequence consistent order. Such record updates are stored at a primary site in a cache memory of a storage controller. A circular queue is created having a head pointer for locating a first (or oldest) record update with subsequent record updates of a track being linked therefrom in sequence consistent order. A tail pointer of the circular queue points to the last record update. A counter field is provided in a track slot header for counting a number of updated records in the track and hence the circular queue. When a data mover in the host processor at the primary site reads record updates for transmission to the secondary site, the data mover starts with the record update pointed to by the head pointer such that the data mover moves through the circular queue (and hence the record updates) in a backward chain fashion. The counter is decremented as each record update in the track is read by the data mover such that when the counter reaches zero the data mover knows that no other record updates exist for that track.

摘要翻译： 数据存储系统通过以顺序一致的顺序将记录更新异步发送到辅助站点来提供灾难恢复能力。 这样的记录更新被存储在存储控制器的高速缓冲存储器中的主站点处。 创建具有用于定位第一（或最早）记录更新的头指针的循环队列，以及以顺序一致的顺序从其链接的轨道的随后记录更新。 循环队列的尾部指针指向最后一个记录更新。 一个计数器字段被提供在轨道槽头中，用于对轨道中的更新记录数进行计数，因此对循环队列进行计数。 当主站点的主机处理器中的数据移动器读取记录更新以传送到辅助站点时，数据移动器从头指针指向的记录更新开始，使得数据移动器移动通过循环队列（并且因此 记录更新）。 当数据移动器读取轨道中的每个记录更新时，计数器递减，使得当计数器达到零时，数据移动器知道该轨道不存在其他记录更新。

公开(公告)号：US06408348B1

公开(公告)日：2002-06-18

申请号：US09372985

申请日：1999-08-20

申请人： Lawrence Carter Blount ,  Emeline Yvonne Germain ,  Vernon John Legvold ,  William Frank Micka ,  Warren Keith Stanley ,  Stephen Charles West

发明人： Lawrence Carter Blount ,  Emeline Yvonne Germain ,  Vernon John Legvold ,  William Frank Micka ,  Warren Keith Stanley ,  Stephen Charles West

IPC分类号： G06F1300

CPC分类号： G06F13/385

摘要： Disclosed is a system, method, and program for limiting input/output (I/O) requests from an application program to a storage device to facilitate the processing of data in memory. A determination is made as to whether an amount of a set of data in memory exceeds at least one threshold. At least one of multiple paths to the storage device is indicated as blocked to I/O requests from the application program. Each indicated blocked path is available for processing the set of data in memory. An I/O request from the application program is received and a determination is made as to whether the received I/O request intends to access the storage device using one indicated blocked path. The I/O request is denied access to the storage device via the blocked path.

摘要翻译： 公开了一种用于将来自应用程序的输入/输出（I / O）请求限制到存储设备以便于对存储器中的数据进行处理的系统，方法和程序。 确定存储器中的一组数据的量是否超过至少一个阈值。 到存储设备的多个路径中的至少一个被指示为来自应用程序的I / O请求的阻塞。 每个指示的阻塞路径可用于处理存储器中的数据集。 接收来自应用程序的I / O请求，并确定所接收的I / O请求是否打算使用一个指示的阻止路径访问存储设备。 I / O请求被拒绝通过阻塞路径访问存储设备。

公开(公告)号：US5991775A

公开(公告)日：1999-11-23

申请号：US949669

申请日：1992-09-23

申请人： Brent Cameron Beardsley ,  Lawrence Carter Blount ,  Joel Harvey Cord ,  Michael Howard Hartung ,  Vernon John Legvold

发明人： Brent Cameron Beardsley ,  Lawrence Carter Blount ,  Joel Harvey Cord ,  Michael Howard Hartung ,  Vernon John Legvold

IPC分类号： G06F12/08 ,  G06F12/12 ,  G06F12/02

CPC分类号： G06F12/0866 ,  G06F12/123 ,  G06F12/084 ,  G06F2212/312 ,  Y10S707/99956

摘要： A data storage system provides generalized record caching through a control unit adapted to support track caching in the upper level store of a two level memory. Dynamic reallocation of space between each type of caching in the upper store follows operating patterns of host computer systems using the data storage system. A storage controller cache has a plurality of segments. A directory data entry data structure is allocated each segment. Such allocated directory entries are used to identify tracks as cached. A plurality of unallocated directory entries are also provided. As a record is cached in a segment outside of a track slot, an unallocated directory entry is used to identify a virtual track in cache corresponding to the track of the record in the lower level store. Records from one track can thus appear in several segments outside track slots. Tracking of records to locate records least recently used is done globally over all track slots and record caching segments and locally within individual record caching segments. A mechanism is provided for identifying record slots, as they become least recently used, and dropping them from the upper level store in the face of competing demands for the space. A second mechanism identifies least recently used segments for dropping from the upper level store.

摘要翻译： 数据存储系统通过适于支持两级存储器的上层存储器中的轨道缓存的控制单元提供广义记录高速缓存。 在上层商店中的每种类型的缓存之间的动态重新分配空间遵循使用数据存储系统的主机系统的操作模式。 存储控制器高速缓存具有多个段。 每个段分配目录数据条目数据结构。 这样分配的目录条目用于将轨道标识为缓存。 还提供了多个未分配的目录条目。 由于记录被缓存在轨道槽外的段中，所以使用未分配的目录条目来标识与低级存储中的记录的轨道相对应的高速缓存中的虚拟轨道。 因此，来自一个轨道的记录可以出现在轨道槽外部的若干段中。 跟踪记录以定位最近最少使用的记录是在所有轨道插槽上全局完成的，并记录缓存段，并在本地记录缓存段。 提供了一种用于识别记录槽的机制，因为它们最近被使用，并且在面对空间的竞争性需求的情况下将其从高级存储中删除。 第二种机制识别最近最少使用的段从高级存储中删除。

公开(公告)号：US5802547A

公开(公告)日：1998-09-01

申请号：US729899

申请日：1996-10-15

申请人： Vernon John Legvold

发明人： Vernon John Legvold

IPC分类号： G06F12/08 ,  G06F13/12 ,  G06F12/00 ,  G06F13/00

CPC分类号： G06F13/122 ,  G06F12/0866

摘要： A data storage system establishes streamlined data exchange paths between stored data and multiple system hosts, decreasing contention among the hosts and expediting host access to the data. First, one or more storage directors receive an initial host request to access a data block contained in part of a shared data resource. This part of the shared data resource may comprise a magnetic disk drive device, part of a physical device, a logical volume, a logical device or another data group. In response, the storage director performs a predetermined series of initial access steps to facilitate the data transfer. For example, the data block is copied from the shared data resource into a cache, if not already present therein. Then, a data path is formed between the host and the device containing the data block. Next, a fast path map is updated to indicate establishment of the data path between the host and the device containing the data block. The data block is subsequently exchanged between the cache and the host. When the storage director receives a subsequent request by the same host to access the same device, the storage director consults the fast path map and determines that a data path already exists between the host and the device. Accordingly, the requested data is exchanged between the host and the cache utilizing the previously established data path between the host and the device, avoiding many steps initially required to establish the data path.

摘要翻译： 数据存储系统在存储的数据和多个系统主机之间建立简化的数据交换路径，减少主机之间的争用，并加快主机对数据的访问。 首先，一个或多个存储引导器接收初始主机请求以访问包含在共享数据资源的一部分中的数据块。 共享数据资源的这一部分可以包括磁盘驱动器设备，物理设备的一部分，逻辑卷，逻辑设备或另一数据组。 作为响应，存储指导器执行预定的一系列初始访问步骤以便于数据传送。 例如，数据块从共享数据资源复制到高速缓存（如果尚未存在）。 然后，在主机和包含数据块的设备之间形成数据路径。 接下来，更新快速路径图以指示主机和包含数据块的设备之间的数据路径的建立。 数据块随后在缓存和主机之间交换。 当存储主管接收到相同主机访问相同设备的后续请求时，存储主管查阅快速路径映射并确定主机与设备之间已经存在数据路径。 因此，使用主机和设备之间先前建立的数据路径，在主机和高速缓存之间交换所请求的数据，避免了建立数据路径最初需要的许多步骤。