公开(公告)号：CA2761553C

公开(公告)日：2019-03-05

申请号：CA2761553

申请日：2011-12-09

Applicant: IBM CANADA LTD IBM CANADA LIMITEE

Inventor： SACHEDINA AAMER ,  HURAS MATTHEW A

IPC: G06F12/00 ,  G06F12/0866

Abstract: An illustrative embodiment of a computer-implemented process for logical buffer pool extension identifies a page in a memory for eviction, analyzes characteristics of the page to form a differentiated page, wherein the characteristics of the page comprise descriptors including a workload type, a page weight, a page type, frequency of access and timing of most recent access and identifies a target location for the differentiated page from a set of locations including a fastcache storage and a hard disk storage to form an identified target location. The process further selects an eviction operation from a set of eviction operations using the characteristics of the differentiated page and the identified target location and writes the differentiated page to the identified target location using the selected eviction operation wherein the differentiated page is written only to the fastcache storage.

公开(公告)号：CA2427071C

公开(公告)日：2010-07-06

申请号：CA2427071

申请日：2003-04-28

Applicant: IBM CANADA

Inventor： CRANSTON LESLIE A ,  BHATTACHARJEE BISHWARANJAN ,  HURAS MATTHEW A ,  LINDSAY BRUCE G ,  MALKEMUS TIMOTHY R ,  MCARTHUR CATHERINE S ,  PADMANABHAN SRIRAM K ,  WINER MICHAEL J

IPC: G06F16/22

Abstract: An information retrieval system and method of operation therefor, wherein a plurality of dimensions for a table are identified, and at least one block in a plurality of blocks in the table is associated with a dimension value for each dimension in the plurality of dimensions, where each block comprises a plurality of contiguous storage pages. Additionally, the at least one block can be further associated with an associated cell, wherein the associated cell has a unique combination of dimension values comprising an associated dimension value for each dimension in the plurality of dimensions. A unique associated bit list for each dimension value for each dimension in the plurality of dimensions has a unique corresponding list entry for each block associated with that dimension value, and a unique associated bit list for each cell has a unique corresponding list entry for each block associated with that cell. These unique associated bit lists facilitate searching, insertion and deletion of records. Each list entry in the unique associated bit lists may also have an associated storage space indicia bit for its associated block for indicating whether the block is likely to be full for further facilitating insertion of records and management of data storage space. Each block may also have a unique associated empty page bitmap for indicating when pages in the block are empty and for identifying when the entire block is empty so that the block may be disassociated from its associated dimension values and associated cell.

公开(公告)号：CA2384185A1

公开(公告)日：2003-10-29

申请号：CA2384185

申请日：2002-04-29

Applicant: IBM CANADA

Inventor： ROMANUFA KERILEY K ,  SACHEDINA AAMER ,  HURAS MATTHEW A

IPC: G06F12/08 ,  G06F12/00 ,  G06F17/30 ,  G06F12/02 ,  G06F9/44

Abstract: A hash table is provided in which identifying properties of the data items pointed to by the hash table are selected and are included within the hash table buckets. The properties recorded in the buckets are able to be stored in a cache for the hash table and are used to located data items pointed to by the hash table. The hash table has bucket groups defining a second level hash table. The bucket groups are used to permit resizing of the hash table. Each bucket group has a latch so that that exclusive access for the resizing operation is obtainable on a group by group basis.

公开(公告)号：CA2377649A1

公开(公告)日：2003-09-20

申请号：CA2377649

申请日：2002-03-20

Applicant: IBM CANADA

Inventor： WILDING MARK F ,  HURAS MATTHEW A

IPC: G06F17/30 ,  G06F15/16 ,  G06F9/44

Abstract: A distributed database management system implemented on a cluster of computers. The cluster including disks organized as data nodes. The data nodes being locall y connected to one or more computers in the cluster. The distributed database management system includes a distributed database manager layer having a lock manager, a resource manager , and an architecture manager. The lock manager serializes access to the data nodes i n the cluster. The resource manager maintains a data representation of the resources of each computer in the cluster and makes the data available to other computers in the cluster to facilitate access to the resources by the other computers. The architecture manager defines logical ownership relationships between data nodes and computers in the cluster to permit the architecture o f the database to be dynamically reconfigured to provide improved performance.

公开(公告)号：CA2281287C

公开(公告)日：2003-03-11

申请号：CA2281287

申请日：1999-09-01

Applicant: IBM CANADA

Inventor： HOP HING NELSON ,  WINER MICHAEL J ,  HURAS MATTHEW A ,  LINDSAY BRUCE G ,  CRANSTON LESLIE A

IPC: G06F17/30

Abstract: A method of searching a table in a database management system for free space for inserting a new record into the table. The table resides in a storage media, and includes a series of pages capable of storing a plurality of records. The table has a clustering index defined in a column of the table. The new record has a new record attribute for storing in the column o f the table when the new record is stored in the table. The method comprises searching for a target page in the series of pages, the target page being selected to maintain clustering when the new record is inserted thereon. Once the target page has been located, it is searched for sufficient free space t o accommodate the new record. If sufficient free space is found on the target page, then the new record is inserted on the target page. If sufficient free space is not found on the target page, then a target neighbourhood of pages in the series of pages is searched for sufficient free space to accommodate the new record. The target neighbourhood of pages surrounds the target page. If sufficient free space is found on a page in a target neighbourhood of pages, then the new record is inserted on to this page in the target neighbourhood of pages. If sufficient free space is not found in the target neighbourhood of pages, then the methodfurther comprises searching for sufficient free space to accommodate the new record outside the target neighbourhood of pages. If sufficient free space is found outside the target neighbourhood of pages, then the new record is inserted on a non-neighbourin g page outside the target neighbourhood of pages. If sufficient free space is not found outside the target neighbourhood of pages, then a new page having the new record is appended to the end of table.

公开(公告)号：CA2353017A1

公开(公告)日：2003-01-12

申请号：CA2353017

申请日：2001-07-12

Applicant: IBM CANADA

Inventor： LI JOHN ,  HURAS MATTHEW A ,  SACHEDINA AAMER ,  ROMANUFA KERILEY K ,  SCHLAMB KELLY W ,  ZHENG ROGER L

IPC: G06F3/06 ,  G06F11/14 ,  G06F11/20 ,  G06F17/30 ,  G11B27/10 ,  G11B25/04

公开(公告)号：CA2146170A1

公开(公告)日：1996-10-04

申请号：CA2146170

申请日：1995-04-03

Applicant: IBM CANADA

Inventor： HURAS MATTHEW A ,  VINCENT TIM J

IPC: G06F9/00 ,  G06F9/46 ,  G06F9/48 ,  G06F15/16 ,  G06F15/163

Abstract: A service provider for use in a client-server system which is capable of detecting the abnormal termination of a client process is disclosed. The service provider does not require a dedicated process for polling client processes in order to verify their status. Rather, a semaphore, which is used in conjunction with a shared memory segment for communication between a client process and the service provider, is initialized in such a manner that the operating system will automatically increment the semaphore in the event the client process is terminated. Thus, the semaphore will be incremented either when the client process deliberately increments the semaphore in order to notify the service provider that the client process has written data to a shared memory segment, or the semaphore will be incremented by the operating system in the event the client process terminates. A test flag is established in shared memory in order to differentiate whether the semaphore was incremented by the client process, or by the operating system. The client process will set the flag only when the client process increments the semaphore. Therefore, whenever the semaphore is incremented, the service provider will test the condition of the flag, and terminate resources allocated to the client process if the flag is not set.

公开(公告)号：CA2433254A1

公开(公告)日：2004-12-25

申请号：CA2433254

申请日：2003-06-25

Applicant: IBM CANADA

Inventor： HURAS MATTHEW A ,  SACHEDINA AAMER ,  ROMANUFA KERILEY K

IPC: G06F12/02 ,  G06F12/08 ,  G06F13/38 ,  G06F17/30 ,  G11C7/10

Abstract: A system and method for warm shutdown and restart of a buffer pool is provided. The buffer content including pages and control structures associated with these pages i s partitioned into blocks and each block is saved to disk as a sequential file. The size of eac h block may be selected to provide the optimal I/O (input/output) efficiency during buffer pool shutdown and restart operations. Blocks can be saved simultaneously using a number of writer threads. During restart of the buffer pool, the control information is read from the disk an d restored to the buffer pool using reader threads. The buffer pool may be made available for references to pages while reader threads are enabled to read the saved information by blocks to facilitate a more rapid restoration of information in the buffer pool. Allocation of portions of the buffer pool for restoring pages may be performed while the buffer pool is available for references to hasten availability.

公开(公告)号：CA2427071A1

公开(公告)日：2004-10-28

申请号：CA2427071

申请日：2003-04-28

Applicant: IBM CANADA

Inventor： LINDSAY BRUCE G ,  BHATTACHARJEE BISHWARANJAN ,  PADMANABHAN SRIRAM K ,  MALKEMUS TIMOTHY R ,  HURAS MATTHEW A ,  MCARTHUR CATHERINE S ,  CRANSTON LESLIE A ,  WINER MICHAEL J

IPC: G06F16/22 ,  G06F17/30

Abstract: An information retrieval system and method of operation therefor, wherein a plurality of dimensions for a table are identified, and at least one block in a plurality of blocks in the table is associated with a dimension value for each dimension in the plurali ty of dimensions, where each block comprises a plurality of contiguous storage pages. Additionally, the at least one block can be further associated with an associated cell, wherein t he associated cell has a unique combination of dimension values comprising an associated dimension value for each dimension in the plurality of dimensions. A unique associated bit list for each dimension value for each dimension in the plurality of dimensions has a uniq ue corresponding list entry for each block associated with that dimension value , and a unique associated bit list for each cell has a unique corresponding list entry for each block associated with that cell. These unique associated bit lists facilitate searching, insertion and deletion of records. Each list entry in the unique associated bit lists may also have an associated storage space indicia bit for its associated block for indicating whether the block is likely to be full for further facilitating insertion of records and management of data storage space. Each block may also have a unique associated empty page bitmap for indicating whe n pages in the block are empty and for identifying when the entire block is empty so that t he block may be disassociated from its associated dimension values and associated cell.

公开(公告)号：CA2426606A1

公开(公告)日：2004-10-25

申请号：CA2426606

申请日：2003-04-25

Applicant: IBM CANADA

Inventor： MCARTHUR CATHERINE S ,  HURAS MATTHEW A ,  THORNBOROUGH SHAUN K ,  WINER MICHAEL J ,  HONG QUANHUA CLAIRE ,  ZHENG ROGER L Q ,  MEADOWCROFT LORYSA M

IPC: G06F16/901 ,  G06F17/30 ,  G06F12/02 ,  G06F17/40

Abstract: The present invention is directed to using a memory buffer to facilitate log catchup for online operations. Some designated areas of memory are allocated for one or more buffers for use during online operations (e.g. reorganization, create index, etc.). Concurrent update activities to the target object write informational records into the one or more buffer s. If the buffers become full, the informational records may be written to logs for persistent storage. After online operations finish building all or a suitable portion of the shadow object, l og catchup is performed by first applying activities in the informational records from the buffer an d thereafter applying activities from logs, if necessary. The invention helps solve the drawbacks of log catchup since the buffer contains only the information related to the target shadow object . Also it does not involve physical I/Os, as these records are all in memory.