Browse Prior Art Database

Structuring Data for Transfer And Sharing in a Distributed Environment

IP.com Disclosure Number: IPCOM000100438D
Original Publication Date: 1990-Apr-01
Included in the Prior Art Database: 2005-Mar-15
Document File: 2 page(s) / 76K

Publishing Venue

IBM

Related People

Bryden, TM: AUTHOR [+4]

Abstract

This article presents a new solution to the problems posed by incompatible data representation formats in an interconnection of heterogeneous computers. In such an environment, groups of cooperating tasks residing on different machines typically interact through some intertask communication facility. These interactions are complicated by incompatible data representations on machines with different architectures. The solution presented here frees tasks from these complications.

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

Structuring Data for Transfer And Sharing in a Distributed Environment

       This article presents a new solution to the problems
posed by incompatible data representation formats in an
interconnection of heterogeneous computers.  In such an environment,
groups of cooperating tasks residing on different machines typically
interact through some intertask communication facility.  These
interactions are complicated by incompatible data representations on
machines with different architectures.  The solution presented here
frees tasks from these complications.

      The key to the solution is to store all data in data units.  A
data unit is a collection of information with a defined structure.
The structure is specified by the task which creates the data unit.
It is known by the underlying operating system, or environment, which
supports the existence of tasks throughout the network.  The
environment is responsible for providing all essential data unit
services, including allocation of new data units.  Since the
environment knows the structure of data units, it can transparently
perform any necessary translations when a task wishes to transfer or
share data.

      All data units consist of a collection of primitive atomic
entities.  A set of atomic entities, or simple data types, is defined
by the environment.  This set includes most, if not all, simple data
types that may be useful to tasks.  The following simple data types
are included in the set:
      Byte
      Character
      Unsigned integer (various sizes)
      Signed integer (various sizes)
      Floating point (various sizes)
      Pointer (restricted)

      A data unit may consist of any one of the simple data types.
However, a task programmer may also wish to combine the simple types
to form more complex data types. Therefore, the data structure's
record and array are also supported.  An array is an indexed sequence
of elements of the same type...