| dc.description.abstract | The widespread use of Distributed Computing Systems (DCSs) is due
to the advantages of better performance, improved reliability, efficient
resource sharing and easy upgradability. In order to realize the
advantages of a DCS, many issues associated with their design should be
tackled efficiently. One such significant issue is the design and
development of a protocol and same is the theme of this thesis. The
development of a protocol typically consists of studies related to its
design, specification, validation and performance evaluation. The
research in the area of protocols has been mainly concentrated on the
specification and validation aspects. Many techniques have been
proposed for these two aspects. But prior to the initiation of the
actions of specification and validation of the protocol, the protocol
should be designed systematically. Not much research work has been
reported in this area. In view of this, an effort has been made in this
thesis to develop a systematic procedure for designing protocols for a
DCS, starting from the specifications of a DCS and the protocol
requirements.
The design and development of a protocol can be divided into six
phases: (i) specification of the DCS, (ii) specification of protocol
requirements, (iii) protocol design, (iv) specification and validation
of the designed protocol, (v) performance evaluation and (vi) hardware/
software implementation. In this thesis a formal method based on matrix
and set theoretic concepts has been developed for the specification of a
DCS keeping in view the protocol design aspects. The DCS is
characterized by various properties such as concurrency, strong
concurrency, exclusiveness and sequencing. A 6-tuple model is
presented to specify a DCS and methods are given to compute these
properties of the DCS from this model. The technique presented for the
specification of a DCS has many advantages as compared to other existing
techniques of specification. Further, inconsistencies of specification
and deadlock caused due to improper specification can easily be detected
using this technique. A similar technique is developed for the
specification of protocol requirements. These two formal specification
techniques form the basis of the development of simple and
straightforward procedures for the design of the protocol. The
applicability of the above design procedure has been illustrated by
considering an example of a computing system encountered onboard a
spacecraft. A Petri net based approach has been adopted to model the
protocol. New methods are developed to compute the invariants of Petri
nets and colored Petri nets. These invariants are used to validate the
protocol. The validation performed on the designed protocol confirms
the liveness, deadlock freeness, 1-boundedness or safeness and recovery
from failure properties of the protocol. For the sake of illustration,
only one failure mode i.e., the corruption of messages at the receiving
end is considered. However, on similar lines other modes of failure can
also be considered in the design. The methodology developed in this
thesis can also be used in other DCS applications such as distributed
industrial control and distributed data processing. | |
