written 8.4 years ago by | • modified 8.4 years ago |
Mumbai University > Information Technology > Sem6 > Distributed System
Marks: Dec 2015
Year: 5M
written 8.4 years ago by | • modified 8.4 years ago |
Mumbai University > Information Technology > Sem6 > Distributed System
Marks: Dec 2015
Year: 5M
written 8.4 years ago by |
The process of converting from a particular machine representation to external data representation format is called serializing, and the reverse process called deserializing.
The standard data representation format is called external data representation, and its designer must successfully handle the problem of different representations for data such as characters, integers, and floating-point numbers. Of these, the handling of floating-point numbers needs special precautions. The issues discussed by Maguire and Smith [1988] for handling floating-point numbers in external data representation schemes are given below.
Fig. 8.5 a) Example illustrating the need for 12 pieces of translation software in a heterogeneous system having 4 types of processors. b) Example illustrating the need of only 8 pieces of translation software in heterogeneous system having 4 types of processors when the external data representation mechanism is used.
A floating point number representation consists of a n exponent part, a mantissa part, and a sign part. The issue of proper handling of the exponent and the mantissa has been described separately because the side effects caused by an external data representation affect each of the two components differently.