nc,
ely
wth
age
ion
'ead
ize
and
cal
rs,
lar
the
| of
lage
as
sfer
over
the
ture
lata
the
sfer
e is
rget
sult
in no loss of data content or distortion of relationships between data
fields.
The first transformation converts the data in the source computer to the
standard Interchange Format, which is conveyed in some manner to the second
machine, where a second transformation regenerates the data in the target
computer system format.
The conversions to and from the mutually agreed upon transport medium
constitute the actual transfer. The medium may be magnetic tape, floppy
disk or other physical media, or electronic data transmission. The transfer
protocol must be mutually agreed upon and common. Inverse conversion
routines must be used for mutual understanding of the conveyed bit patterns.
It is important to note that the data transfer model is not concerned with
information content, which is assumed to be coded into the data.
This model can be compared with the ISO Open Systems Interconnection (OSI)
seven layer model (ISO Standard 7498). In this model, the first five levels
define the data transmission itself, and are not of concern in this
discussion. Level 6, the Presentation Layer, defines the rules for building
an application-independent logical structure of the message and builds the
message for transmission. On the receiving end, it parses the message to
recover the data structure through a knowledge of the Level 6
application-independent structure rules. Level 7, the Applications Layer,
contains a data file incorporating the information required. On the
receiving end, the application-specific processing is performed and the
meanings of the data files are understood. Conversion to and from the
Standard Interchange Format is viewed as a task at the Presentation Layer
(Level 6). With this structure, media considerations are eliminated from
the Standard Interchange Format. Furthermore, discipline/application
independence is achieved by describing the data bases with a standard
syntax.
STANDARDIZATION
Over the years, many attempts have been made to develop standards for the
interchange of data. At some levels, a degree of success has been achieved.
For example, the method of recording data on magnetic computer tape at
various bit densities has industry-wide acceptance. This probably
represents the ONLY universally accepted data transfer standard in
existence.
The problem of developing universally acceptable standards is very complex.
In the absence of an accepted standard, each group wishing to interchange
data defines its own format and/or protocol. Typically, this is quite
application dependent. The objective is usually to transfer data between
two systems with the minimum of development cost. Thus, the format tends to
be as simple as possible, with most of the format information defined
externally. There is usually little flexibility.
This approach is completely acceptable for a closed group, which may in fact
be quite large. Where all members of the group have the same hardware and
software, a simple copy or backup of the computer files using the resident
operating system may be the most efficient method of transferring data. The
computer system resident communication protocol may be used for electronic
data transmission. Problems begin to arise when the group is expanded to
include other computers and different hardware.
183