Abstract
M.Sc. (Computer Science)
In this dissertation we propose extensions to the Self Protecting Object (SPO) model to facilitate the sharing of information in a more effective manner. We see the sharing ofinformation as the sharing of objects that provide services. Sharing objects effectively is allowing the objects to be used in a secure environment, independent of their location, in a manner usage was intended. The SPO model proposed by Olivier [32] allows for objects in a federated database to be moved from one site to another and ensures that the security policy of the object will always be respected and implemented, regardless of its location.
Although the SPO model does indeed allow for objects (information) to be shared
effectively, it fails to address issues of maintaining integrity within objects. We therefore define the notion of maintaining integrity within the spa model and propose a model to achieve it. We argue that ensuring an SPO is only used in a way usage was intended does not suffice to ensure integrity. The model we propose is based on ensuring that modifications to an SPO are only executed if the modification does not violate the constraints defined for the Sf'O, The model" allows for an spa to maintain its unique identity in addition to maintaining its integrity.
The SPO model is designed to be used in a federated database on sites that are
stationary. Therefore, having addressed the issue of maintaining integrity within
SPOs on stationary sites in the federated database, we then introduce the notion
of a mobile site: a site that will eventually disconnect from the federated database
and become unreachable for some time.
Introducing the mobile site into the federated database allows us to propose the
Mobile Self Protecting Object (MSPO) and its associated architecture. Because
of the nature of mobile sites, the original model for maintaining integrity can not
be applied to the MSPO architecture. We therefore propose a mechanism (to be
implemented in unison with the original model) to ensure the integrity of MSPOs
on mobile sites.
We then discuss the JASPO prototype. The aim of the prototype was to determine
if the Self Protecting Object model was feasible using current development technologies. We examine the requirements identified in order for the prototype
to be successful and discuss how these were satisfied. Several modifications were
made to the original spa model, including the addition of a new module and the
exclusion of others, we discuss these modifications and examine why they were necessary.