-
公开(公告)号:JPS60147855A
公开(公告)日:1985-08-03
申请号:JP24355384
申请日:1984-11-20
Applicant: IBM
Abstract: © A storage-management system executable in one host or simultaneously in a plurality of hosts (10, 11), manages concurrently executing migration (28), recall (27), and recycle (30) (such as defragmentation) tasks. Accesses to data storage volumes (14-18) (tape, disks, etc.) are managed in accordance with predetermined contention resolution procedures and priorities. Certain tasks are non-interruptible.
-
公开(公告)号:DE69225636D1
公开(公告)日:1998-07-02
申请号:DE69225636
申请日:1992-03-31
Applicant: IBM
Inventor: BEGLIN THOMAS WILLIAM
Abstract: The average time a user must wait to have an object retrieved from an automated optical disk library is reduced by a method for prioritizing read requests. When a read request is received it is added to a queue of pending requests. All pending requests associated with volumes currently mounted on optical disk drives are processed first. The pending requests associated with the opposite sides of the currently mounted volumes are processed next. The next requests to be processed are those pending read requests associated with the unmounted volume having the greatest number of pending read requests. Thereafter any other pending requests for that unmounted volume will be processed as normally would be done for a currently mounted volume. This method continues until all pending read requests have been processed for the unmounted volumes. Allowing pending requests to go unprocessed, a problem referred to as starvation, is avoided by determining a mean response time of processed requests and increasing the priority of pending requests associated with unmounted volumes having a response time greater than some factor of the mean response time.
-
公开(公告)号:DE3587398D1
公开(公告)日:1993-07-22
申请号:DE3587398
申请日:1985-01-02
Applicant: IBM
Abstract: © A storage-management system executable in one host or simultaneously in a plurality of hosts (10, 11), manages concurrently executing migration (28), recall (27), and recycle (30) (such as defragmentation) tasks. Accesses to data storage volumes (14-18) (tape, disks, etc.) are managed in accordance with predetermined contention resolution procedures and priorities. Certain tasks are non-interruptible.
-
公开(公告)号:DE69225636T2
公开(公告)日:1999-02-25
申请号:DE69225636
申请日:1992-03-31
Applicant: IBM
Inventor: BEGLIN THOMAS WILLIAM
Abstract: The average time a user must wait to have an object retrieved from an automated optical disk library is reduced by a method for prioritizing read requests. When a read request is received it is added to a queue of pending requests. All pending requests associated with volumes currently mounted on optical disk drives are processed first. The pending requests associated with the opposite sides of the currently mounted volumes are processed next. The next requests to be processed are those pending read requests associated with the unmounted volume having the greatest number of pending read requests. Thereafter any other pending requests for that unmounted volume will be processed as normally would be done for a currently mounted volume. This method continues until all pending read requests have been processed for the unmounted volumes. Allowing pending requests to go unprocessed, a problem referred to as starvation, is avoided by determining a mean response time of processed requests and increasing the priority of pending requests associated with unmounted volumes having a response time greater than some factor of the mean response time.
-
公开(公告)号:DE3587398T2
公开(公告)日:1993-12-09
申请号:DE3587398
申请日:1985-01-02
Applicant: IBM
Abstract: © A storage-management system executable in one host or simultaneously in a plurality of hosts (10, 11), manages concurrently executing migration (28), recall (27), and recycle (30) (such as defragmentation) tasks. Accesses to data storage volumes (14-18) (tape, disks, etc.) are managed in accordance with predetermined contention resolution procedures and priorities. Certain tasks are non-interruptible.
-
-
-
-