公开(公告)号：CA2322603A1

公开(公告)日：2002-04-06

申请号：CA2322603

申请日：2000-10-06

Applicant: IBM CANADA

Inventor： GORALWALLA IQBAL A ,  WINER MICHAEL J ,  SHARPE DAVID C ,  HURAS MATTHEW A

IPC: G06F17/30 ,  G06F12/02 ,  G06F9/312

Abstract: An optimization for updatable scrollable cursors in database systems. To ensure that a temporary copy of a data record matches the data record maintained in the database, th e temporary copy is associated with a timestamp included on each page storing the data records o f the database. The timestamp on a page in the database is updated whenever any record stored on the page is updated. The optimization compares the timestamp stored with the temporary record wit h the current timestamp associated with the page on which the data record is stored in the database. If the two timestamps are matching then the temporary record is identical to the databa se table record and there is no need to carry out a comparison of the attribute values of the temporar y record with the record in the database table.

公开(公告)号：CA2306969A1

公开(公告)日：2001-10-28

申请号：CA2306969

申请日：2000-04-28

Applicant: IBM CANADA

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

IPC: G06F9/50 ,  G06F9/46 ,  G06F9/48

Abstract: A computer system dynamically scales back latch requests for system resource s. Tasks seeking access to system resources each dynamically determine the probability that t he task will gain access to the latch relating to a given system resource. Where the task estimates that its probability is below a defined threshold, the task will suspend itself for a defined sleep time. The task dynamically adjusts the length of the sleep time based on the number of times the task enters the suspended state and on the relative changes in the estimated probability that the task will gain access to the resource.

公开(公告)号：CA2719673A1

公开(公告)日：2011-01-18

申请号：CA2719673

申请日：2010-11-05

Applicant: IBM CANADA

Inventor： BAE MYUNG M ,  BUCKLER ANDREW D ,  CHAN YUE FONG Y ,  GARCIA-ARELLANO CHRISTIAN M ,  HURAS MATTHEW A ,  MUELLER MARKUS ,  SACHEDINA AAMER ,  SCHMUCK FRANK B ,  VOLOBUEV YURI

IPC: G06F11/07 ,  G06F15/16

Abstract: An illustrative embodiment of a computer-implemented process for fencing shared cluster resources in event of a possible split-brain, identifies a failing resource of a node within a set of shared resources to form an identified failing resource, fences a subset of the set of shared resources to form a winning subset of shared resources and prevents the identified failing resource from communicating with the winning subset of shared resources using a white list maintained at each database system of the winning subset of shared resources. The computer-implemented process further determines whether the identified failing resource has been cleared to form a cleared failing resource and responsive to a determination that the identified failing resource has been cleared, rejoins the cleared failing resource with the winning subset of the set of shared resources absent a re-boot of the cleared failing resource.

公开(公告)号：CA2369621C

公开(公告)日：2009-06-09

申请号：CA2369621

申请日：2002-01-25

Applicant: IBM CANADA

Inventor： WINER MICHAEL J ,  ZHENG ROGER L Q ,  HURAS MATTHEW A ,  GOSS JEFFREY J ,  FLASZA MIROSLAW A ,  POSNER SARAH ,  MEADOWCROFT LORYSA M ,  MCARTHUR CATHERINE S

IPC: G06F9/50 ,  G06F9/46 ,  G06F12/00 ,  G06F12/14 ,  G06F13/14 ,  G06F17/30

Abstract: The present invention is directed at a method and apparatus for handling resource transaction requests. By marking a waiting transaction request, having an access mode which is incompatible with the access mode of at least one of the executing transactions, with an ignore bit, a subsequent transaction request with an access mode compatible with the access mode(s) o f the executing transaction(s) may receive access to the resource. The subsequent transactio n request is then marked with a bypass indicator such that when the executing transactions are finish ed accessing the resource, the waiting transaction request may be notified it may have access to the resource. By including the ignore bit, each subsequent transaction request which arrives after ignore b it has been set, does not suffer from starvation.

公开(公告)号：CA2322604C

公开(公告)日：2005-06-28

申请号：CA2322604

申请日：2000-10-06

Applicant: IBM CANADA

Inventor： XUE XUN ,  ZHENG ROGER L ,  VINCENT TIMOTHY J ,  MCKEOUGH SEAN W ,  HURAS MATTHEW A

IPC: G06F9/46 ,  G06F17/30

Abstract: A deadlock management system for a demultiplexed database system having work er agents and applications associated with and disassociated from the worker agents. Applications holding locks on database resources and capable of retaining locks when disassociated from worker agents. Providing each application with a flag set in accordance with the existing value of the application flag and in accordance with the state of applications holding and requesting locks. The deadlock management system providing for applications to be placed on a wait queue, a priority queue, or provided worker agents or overflow worker agents depending on the flag value for the application. Declaring deadlocks based on the flag value of applications and on the state of the lock wait graph showing applications holding and requesting locks.

公开(公告)号：CA2322603C

公开(公告)日：2005-04-12

申请号：CA2322603

申请日：2000-10-06

Applicant: IBM CANADA

Inventor： HURAS MATTHEW A ,  WINER MICHAEL J ,  GORALWALLA IQBAL A ,  SHARPE DAVID C

IPC: G06F17/30 ,  G06F12/02 ,  G06F9/312

Abstract: An optimization for updatable scrollable cursors in database systems. To ensure that a temporary copy of a data record matches the data record maintained in the database, th e temporary copy is associated with a timestamp included on each page storing the data records o f the database. The timestamp on a page in the database is updated whenever any record stored on the page is updated. The optimization compares the timestamp stored with the temporary record wit h the current timestamp associated with the page on which the data record is stored in the database. If the two timestamps are matching then the temporary record is identical to the databa se table record and there is no need to carry out a comparison of the attribute values of the temporar y record with the record in the database table.

公开(公告)号：CA2426640A1

公开(公告)日：2004-10-28

申请号：CA2426640

申请日：2003-04-28

Applicant: IBM CANADA

Inventor： HELLERSTEIN JOSEPH L ,  ROSE KEVIN R ,  LIGHTSTONE SAM S ,  PAREKH SUJAY S ,  HURAS MATTHEW A

IPC: G06F9/44 ,  G06F11/30 ,  G06F11/34

Abstract: An adaptive throttling system for minimizing the impact of non-producti on work on production work in a computer system. The adaptive throttling system throttles production work and non-production work to optimize production. Th e adaptive throttling system allows system administrators to specify a quantified limit on the performance impact of non-production or utility work on production work. The throttling rate of the utility is then automatically determined by a supervisory agent, so that the utilities' impact is kept within the specifie d limit. The adaptive throttling system adapts dynamically to changes in workloads so as to ensure that valuable system resources are well utilized and utility work is not delayed unnecessarily.

公开(公告)号：CA2302981C

公开(公告)日：2004-06-01

申请号：CA2302981

申请日：2000-03-29

Applicant: IBM CANADA

Inventor： GOSS JEFFREY J ,  HURAS MATTHEW A ,  LINDSAY BRUCE G ,  HOP HING NELSON

IPC: G06F17/30 ,  G06F7/24

Abstract: A database table reorganization is defined to permit online access of the table during the reorganization. Records are reorganized in the database table by vacating records from a defined number of pages and then filling the pages with records in accordance with a desired ordering for the records. Temporary pointers to the new locations of moved records are us ed to prevent table scanner access to the database table from missing or duplicating records whi le scanning the database table during reorganization. Removal of the temporary pointers is synchroniz ed with the completion of scanning of all table scanners that are commenced during a time when records are being moved as part of a vacating or filling step.

公开(公告)号：CA2387200A1

公开(公告)日：2003-11-22

申请号：CA2387200

申请日：2002-05-22

Applicant: IBM CANADA

Inventor： HURAS MATTHEW A

IPC: G06F9/06 ,  G06F9/40 ,  G06F12/00 ,  G06F12/02

Abstract: One of the primary difficulties that result from using static variables in multi-threaded computer programs is that changes to a static variable made by one thread wi ll be seen by all other threads operating within the same process. Multiple threads cannot use static variables separately because other threads within the process can overwrite the values stored at the variable memory location. Thus, the development of multi-threaded programs using stat ic variables often requires explicit thread harmonization by the programmer. Another problem is that threads within the same process must use unique static variable IDs to avoid reading or writing to the location of another static variable. This also requires thread harmonization by the programmer. Accordingly, in view of the shortcomings associated with existing thread- static data implementations, there remains a need for an efficient thread-static data implementation that can be used on most modern operating systems. The present invention provides embodiments of a method, data processing system, and computer program product that efficientl y implements thread-static data without registering pointers. The present invention efficiently implements static thread data in a multi-threaded program. The static thread data of ea ch thread is allocated tothe top of the virtual memory block belonging to the c orresponding stack of each thread. An address mask is created for each thread so as to extract the base address of the static thread data region of the stack for each thread from the corresponding stack pointer. Wh en a thread in execution needs to access its static thread data, it accesses the static thread data variable according to the virtual memory address extracted from the stack pointer usi ng the address mask.

公开(公告)号：CA2249080A1

公开(公告)日：2000-03-25

申请号：CA2249080

申请日：1998-09-25

Applicant: IBM CANADA

Inventor： WINER MICHAEL J ,  HOP HING NELSON ,  HURAS MATTHEW A ,  LINDSAY BRUCE G ,  ROMANUFA KERILEY K ,  GOSS JEFFREY J

IPC: G06F17/30

Abstract: A method for efficiently searching free space in a relational database management system. The method limits the search to a finite number of space map pages in the free space map. If the configured number of space map pages in the free space map are searched and a page with free space is not found for the row, the row is inserted on the last page, or if no space is available on the last page a new page is created and the row is inserted on the new page. New rows are then appended until some predefined amount of space is filled before a search is done again. As a result, insertion of a row into the database management system does not incur the worst-case cost of searching the entire free space map.