As of 1965 all data storage access strategies depended on
prior knowledge of user application intentions.
Many government agencies were collecting large volumes of data,
but had no idea of how applications might use the accumulated data.
The Advanced Research Projects Agency, ARPA, reasoned that
information contained in the data relationships
was intrinsic to the data itself and should
not be diminished by a computer representation.
ARPA: Data Relationships Are Independent of Data Representations.
An application requesting data should receive the same data independent of
how the data is physically represented on any specific machine.
Therefore, any proposed machine-independent data structure
must preserve data relationships
in such a way that they can be shared by any requesting application.
Machine-Independent Data Structure
In 1965 ARPA initiated the
CONCOMP Project to investigate
the conversational use of computers.
One research objective was to define and implement
a machine-independent data structure.
A definition was obvious, an implementation was not.
Machine-Independent Data Structure:
Any implementation for representing data relationships
that can be accessed and manipulated by applications
without any reliance on how the data relationships are
represented and preserved.
The primary purpose of a data structure
is to provide applications access to preserved data.
There are three major capabilities of a data structure:
1) Support ability for applications to manipulate abstract representations
of data relationships.
2) Support ability of system to manage physical representations of
3) Mapping strategies for embedding and
extracting application data in and from storage data.
For a data structure to be machine-independent
all data access reliance on physical properties of the stored
data must reside on the storage side of the mapping interface.
If any data access strategy requires knowledge of storage organization,
on the application side of the mapping interface,
the data structure is physically-dependent.
Set-Theoretic Data Structure
By 1968 a Set-Theoretic Data Structure, STDS, (satisfying the conditions of a machine-independent data structure) had been implemented at the University of Michigan.
The architecture of STDS uses structured sets to represent application data,
storage data, and to support I/O mappings.
Structured Sets are
with two membership conditions,
as defined under the axioms of Extended Set
STDS capabilities are:
1) Set operations for manipulating application data as sets.
2) Set operations for manipulating storage data as sets.
3) All data I/O performed using set operations.
implementation only supported data represented as
labeled arrays, but it demonstrated the feasibility of a
machine-independent data structure.
documents and semi-structured data can be represented by structured sets,
and are now supported by STDS libraries.
STDS libraries are available to any interested in
using set-accessing I/O capabilities to access Big Data.
Copyright © 2017
INTEGRATED INFORMATION SYSTEMS
« Last modified on 08/07/2017 »
- XSP TECHNOLOGY -