Kia Makki

Towards the design and development of a new architecture for Geographic Information Systems

Existing Geographic Information Systems lack many important features such as the ability to model the real world more adequately and facilities for logical deduction and geometric computation. In this paper, we study the structure and behavior of existing Geographic Information Systems, isolate their drawbacks and explore the applicability of object oriented design, logical deduction and hypermedia to developing better and more efficient geographic information systems. In particular, we focus on the concept, design and development of an architecture for the next generation geographic information systems, which will be able to access, synthesize and reason about large volumes of geographic information more efficiently and effectively, without sacrificing functionality, extensibility and consistency. Of significance is the integration of many diverse forms of technological advancements, including object oriented design, expert systems and multi-media systems.

An ECA object service to support active distributed objects

Abstract Existing Object Management Group (OMG) standards (such as the Common Object Request Broker Architecture specification) provide specific details on how to model distributed objects through the use of an object request broker. However, the specification provided by OMG is not sufficient. For example, it does not specify how active, distributed objects can perform certain functions in response to certain events. It also does not provide any details on how to implement an object service to support active, distributed objects. Such an object service can provide facilities for applications to define events and rules. In view of this, this paper identifies, explores, and provides an approach to model applications as active, distributed objects. The research specifically involves the design, development, and experimental demonstration of an object service based on an active, distributed object model. This object service allows applications to define events and rules. When an event occurs, the object service locates the rule objects and triggers them. It also provides transparency for the objects which generate events. To provide a specific context for the approach, an object service using the specifications provided by the OMG is designed and implemented. The class of application environments for which this approach is suitable and useful includes time-dependent applications and systems that require some form of action such as medical and geographic information systems, distributed work-flow management, distributed expert systems, and computer-integrated manufacturing.

Towards a framework for integrating multilevel secure models and temporal data models

Within many organizations the number of databases containing classified or otherwise sensitive data is increasing rapidly. Access to these databases must be restricted and controlled to limit the unauthorized disclosure, or malicious modification of data contained in them. However, the conventional models of authorization that have been designed for database systems supporting the hierarchical, network and relational models of data do not provide adequate mechanisms to support controlled access of temporal objects and context based temporal information. In this paper we extend the multilevel secure relational model to capture the functionality required of a temporal database, i.e. a database that supports some aspect of time, not counting user-defined time. In particular we assign class access to bitemporal timestamped attributes, and give explicit security classifications to temporal elements.

A new storage organization for temporal data bases

A temporal data base needs to maintain large amounts of data, which causes long delays in accessing information. In this article, we design and develop an efficient storage structure for maintaining both current and historical data. Our strategy consists of two storage organizations, one for current versions and the other for history versions. The organization for current versions follows an efficient storage organization of a conventional nontemporal data base. However, for maintaining the history versions of each item in the data base, we use an extended version of a Fibonacci heap, referred to as a history version heap. The novelty of our approach resides in its ability to insert a new version for an item into the data base in O(1) amortized time and retrieve the oldest version of an item in O(1) amortized time. These bounds are irrespective of the number of existing versions for that item. In addition, we can access or delete any version of a given item in logarithmic amortized time. Because our approach treats the current store and history store differently, it uses different access methods for each store possible.

Efficient multicast tree algorithm in ATM networks

Many applications of ATM networks require the multicast function, which sends the same packet to a group of destinations. This paper proposes an algorithm for multicast routing in ATM networks. We define the network cost of a routing as the combination of the cost of establishing connections, the cost of overall bandwidth and the cost of overall switchings. Our routing algorithm works on the original topology of ATM networks with physical switches and physical links, and considers the different switching functionalities of VP (virtual path) switches and VC (virtual channel) switches in the networks. It generates an optimal multicast routing with the minimal overall network cost. Simulations have been done to compare the quality of the routing generated by our algorithm with those of other major algorithms.

An Efficient Solution to the Critical Section Problem

In a distributed system of N sites an efficient solution to the critical section problem is proposed to allow multiple simultaneous entries to the Critical Section. The proposed solution is token-based, deadlock-free and free from starvation. In all cases except during the book keeping cycle of the algorithm, the number of message exchanges per Critical Section execution is less than or equal to 5. Only during the book keeping cycle which is not done very often the number of message exchanges per Critical Section execution reaches to at most N.

Separating semantics from representation in a temporal object database domain

In this paper, we show how existing notions of temporal data have to be modified because of the transition from the relational to the object model. In particular, we focus on the semantics of time in the context of object databases, and on the identification and definition of a set of generic temporal modeling concepts to capture aspects of temporal object database design, that are not amenable to description by current object database models. These notions provide a richer and more flexible framework for defining temporal objects with roles and for developing temporal object models. As such, they may be regarded as a set of proposed temporal requirements for a temporal object model.

Detection and resolution of deadlocks in distributed database systems

Deadlock detection and resolution is one of the major component of a successful distributed database management system. In this paper, we discuss deadlock detection and resolution strategies and present two approaches for detecting and resolving deadlocks in both general distributed database systems and in distributed real-time database systems. Our first approach is to collect information on connectivity of nodes of the overall Transaction Wait-For Graph (TWFG) of the distributed database system and then use these connectivities information to build a local TWFG at each node of the overall TWGF. We then detect the deadlocks by locating the cycles in each local TWFG. To resolve the deadlocks the nodes involved in those cycles in each local T WFG, are removed until there is no cycle in the local TWFGS. Our second approach continuously checks for the occurrences of a deadlock between different transaction trees. As soon as it detects a deadlock it resolves it by aborting one of the transaction tree which has been initiated more recently. Some of the advantages of our approaches over the approaches which are using Probe messages are: (1) no extra storage required to store different probe messages, (2) no false (Phantom) deadlocks are reported, (3) detects and resolve all deadlocks. In addition, our approaches use less messages and time to detect and resolve all deadlocks in the existing TWFG of the distributed database system.

A coherent architecture for a temporal object data base management system

Abstract The last decade has witnessed the emergence of practical general-purpose data base management systems (DBMS) as “state of the art” technology. In consequence, these systems are now widely used in a variety of application environments. However, many such systems were designed to capture current data, i.e., a snapshot of reality, but are insufficient for those applications in which past or future data are also required. Even so, there has been little effort to design temporal DBMS, and currently there are no prominent commercial temporal DBMS. This article identifies, explores, and provides an approach to designing a coherent architecture for a temporal object DBMS (TODBMS). It specifically involves the design, development, and implementation of TODBMS. We introduce the basic features of our system and provide an overview of the functionalities supported. In addition, we identify the major components of TODBMS and define the types of services each component should offer, as well as common abstractions and interfaces for these components. Application environments include management and geographic information systems, data base visualization, and robotics.

A modular methodology to verify communication protocols

State explosion has been a major problem for verifying communication protocols and distributed algorithms. In this paper we analyze the sources that cause the state explosion problem and propose a modular approach to verifying communication protocols based on networks of communicating finite state machines (CFSMs). Our proposed method is innovative and its time and space complexity is independent of the number of processes in a system. The new methodology can significantly reduce the time and space needed to verify a number of interesting logical properties, as evidenced by our implementations.