Browse Prior Art Database

Dual Throttle to Limit Server Resource Utilization

IP.com Disclosure Number: IPCOM000113395D
Original Publication Date: 1994-Aug-01
Included in the Prior Art Database: 2005-Mar-27
Document File: 2 page(s) / 80K

Publishing Venue

IBM

Related People

Fruchtman, B: AUTHOR [+4]

Abstract

Disclosed is a "dual throttle" algorithm where two configurable threshold values are used to limit a file backup server's usage of processor and network resources.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 52% of the total text.

Dual Throttle to Limit Server Resource Utilization

      Disclosed is a "dual throttle" algorithm where two configurable
threshold values are used to limit a file backup server's usage of
processor and network resources.

      A file backup server fields requests to backup groups of files
to file repositories on other nodes.  The processing for one of these
requests can be divided into two logical parts which heavily utilize
different resources:

1.  Setting up for the backup operation.  This involves determining
    which files have been requested for backup, evaluating their
    eligibility for backup, building a list or file containing the
    names of the files to be backed up, and starting backup subtasks
    to perform the data movement.  This part of processing is a heavy
    user of processor resources as directory trees are traversed,
    file attributes are examined, file name lists are created, etc.

2.  Performing the backup operation.  This is the actual sending of
    the selected files to one or more file repository nodes.  This
    part of processing is a heavy user of network resources as each
    of the independent subtasks dispatched to perform backups
    competes for communication facilities and network bandwidth.
    They are also competing for network resource with any other node
    also trying to perform backups.

      The server running on the node is also competing for resources
with any other applications running on the node at the time.  Network
and node administrators need a way to limit the resource utilization
on the server nodes in a way that is efficient for their particular
configuration and work patterns.  For example, a company that does
all its data backups on third shift when no other applications are
active may want to allow the server processes to use as much
processor and network resource as needed.  On the other hand if
backups are done at the same time that users are signed on performing
interactive tasks, it would be good to severely restrict the backup
server's use of...