Browse Prior Art Database

Task Creation in a Heterogeneous Distributed Environment

IP.com Disclosure Number: IPCOM000100447D
Original Publication Date: 1990-Apr-01
Included in the Prior Art Database: 2005-Mar-15
Document File: 2 page(s) / 88K

Publishing Venue

IBM

Related People

Banerji, A: AUTHOR [+3]

Abstract

A method is disclosed for creating new tasks in a distributed computing system. A distributed application consists of several concurrent, cooperating tasks, each of which may execute on a different computer. Such applications generally originate as a single task which then creates all other necessary tasks. This article describes a flexible, sophisticated method for creating these additional tasks.

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

Task Creation in a Heterogeneous Distributed Environment

       A method is disclosed for creating new tasks in a
distributed computing system.  A distributed application consists of
several concurrent, cooperating tasks, each of which may execute on a
different computer.  Such applications generally originate as a
single task which then creates all other necessary tasks.  This
article describes a flexible, sophisticated method for creating these
additional tasks.

      With the new method, a task can create a child task without
knowing the identity, or even the CPU architecture, of the target
machine.  The new technique has identical semantics for both local
and remote task creation.  It allows implementation of a load
balancing scheme where new tasks are placed on under-utilized sites.
Furthermore, when creating a remote child task, a parent can make
efficient use of communication facilities.

      The task creation process is part of the support system, or
environment, of the distributed computing facility.  The process has
three steps, each of which is described below.  It assumes that some
inter- task communication mechanism is provided by the environment.

      In the first step of the process, a parent task requests the
creation of a child.  The parent may optionally specify the machine
on which the child is to be created; if no machine is specified, the
environment chooses one.  After the identity of the target machine
has been determined, the environment verifies that the parent is
authorized to create children there.  The environment then creates a
task shell on the target machine.  A task shell possesses most of the
attributes of a normal task.  However, it has no executable code.  As
illustrated in the figure, it is considered to be alive, but dormant.

      In the second step of the task creation process, the child task
is given code to...