Browse Prior Art Database

Priority Based Client Scheduling in AIX Windows

IP.com Disclosure Number: IPCOM000121885D
Original Publication Date: 1991-Oct-01
Included in the Prior Art Database: 2005-Apr-03
Document File: 2 page(s) / 63K

Publishing Venue

IBM

Related People

Desai, P: AUTHOR [+2]

Abstract

The problem is the current X server model services 'n' requests from a client before scheduling the next client. If each of these requests is large, clients like the window manager, which provide the user interface, are not serviced frequently. The result is a jerky and unresponsive user interface. Solution:

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 57% of the total text.

Priority Based Client Scheduling in AIX Windows

      The problem is the current X server model services 'n'
requests from a client before scheduling the next client. If each of
these requests is large, clients like the window manager, which
provide the user interface, are not serviced frequently.  The result
is a jerky and unresponsive user interface.
      Solution:

      The scheduling algorithm in the X server needs to be modified
to allow clients that are 'hogging' the server to be pre-empted in
favor of clients that need to be serviced more often.

      Clients that need to be serviced more often, that is, they are
to have a higher priority than the other clients, should be
identified when they register with the server. The client index of
each of these clients is saved. Registration can occur by various
methods.  For example, a client that selects the SubstructureRedirect
mask as part of the X Window System event processing mechanism can be
identified as the window manager and since this client requires that
it be serviced more often than any of the other clients, its client
index is saved and used in the new scheduling model.

      The following is a brief outline of the current scheduling
algorithm:
      o    The server is normally 'asleep' in the select system call
waiting for clients to request service.  When one or more clients
send the server a request, the server starts processing each of the
clients that require attention.
...