Internet services exhibit the trend to be structured according to a three-tier, and in general multi-tier, system organization. Even though the partitioning of applications into multiple tiers provides the potentialities to achieve high modularity and ?exibility, the multiplicity and diversity of the employed components, and their interdependencies, make reliability a complex issue to tackle. In this book we present protocols ensuring the e-Transaction (exactly-once Transaction) guarantees, a recent formalization of end-to-end reliability properties for multi-tier systems. The presented solutions achieve high scalability and exhibit very limited overhead, revealing particularly attractive for large scale service delivery platforms. From a theoretical standpoint, the presented e-Transaction protocols can cope with purely asynchronous systems, where no assumption on the failure detection accuracy can be guaranteed. Some of the building blocks underlying these fault tolerant protocols can also be used to deal with a more general class of failures, which we refer to as “performance failures", which model situations of reduced system responsiveness due to both crashes and overloads.
In this dissertation, we introduce a proposal for building an alternative group communication service that shifts the multicast support from core routers to end-systems. Our proposal, called Host Based Multicast (HBM), operates at application-level and provides an efficient multi-point data distribution service for one-to-many or many-to-many communications. With this approach end-hosts, dedicated servers and/or border routers automatically self-organize into an overlay distribution topology where data is disseminated. This overlay topology can be composed of both unicast connections and native multicast islands (e.g. within each site). Therefore it offers a group communication service to all hosts, even those located in a site that does not have access for any reason, to native multicast routing. HBM is a centralized solution, where everything, including group membership management and overlay topology creation, is under the control of a single Rendez-vous Point (RP). This book focuses on four key aspects: scalability, robustness, efficiency, and security.
Recent advances in storage technologies and high performance interconnects have made possible in the last years to build, more and more potent storage systems that serve thousands of nodes. Additionally, the hierarchical organization of current Petascale systems and of the envisioned Exascale platforms contributes to an increase of the I/O subsystem latency. In this dissertation we present a novel generic parallel I/O architecture for clusters and supercomputers. Our design is aimed at large-scale parallel architectures with thousands of compute nodes. Besides acting as middleware for existing parallel file systems, our architecture provides on-line virtualization of storage resources. Another objective of this thesis is to factor out the common parallel I/O functionality from clusters and supercomputers in generic modules in order to facilitate porting of scientific applications across these platforms.
One of the major challenges faced by the Internet today is the quality of service (QOS) for the real time (Time Critical) multimedia data. Many protocols have been developed to ensure the end to end seamless delivery of the time critical data. Some of these protocols use labeling methodology in which the data traveling between the nodes on internet is labeled so that it may be prioritized to avoid delay. The other protocols ensure the end to end resource allocation and guarantee the bandwidth for the time critical data. Out of these protocols the point of discussion in this study is IntServe (Integrated Services) and DiffServe (Differentiated Services) that are being tested and analyzed for their performance related to the traditional Best-Effort.
The aim of this book is to develop a Multi-Level security model for web service computing environments. This book describes about an Hardware Level Authentication scheme of web service consumers for secure communication with the service providers and History based Accounting and Reacting Denial of Service attack (DoS) model to detect and prevent possible DoS attacks in web service computing. The main objectives of this book are to provide built in authentication for web service consumers. It also minimizes the security constraints and challenges in any close group web service communication. This book has engaged the E-learning framework as a test bed to carry out the experiments in support of the proposed idea and its goals. The E-learning framework provides all the possible aspects of web service computing to perform the experimentation of this book. The experimental results in this book reveal the performance and accuracy of the Two Tier Security Framework for SOS.
Managing product variety poses several challenges at the manufacturer’s end in terms of inventory control. Many manufacturers incur increased inventory costs (raw material, in process and finished goods) to guarantee high quality of service to customers. Many companies seeking the right balance of product variety, flexibility and costs choose to manufacture products in a batch production mode. However, batch production systems involve complex trade-offs between productivity losses resulting from making too many small batches and the costs of tying up working capital in inventory as large batches. These trade-offs gets magnified as product variety increases. The objective of this research is to develop a simple analytical model for capacitated multi-product batch production systems that could be used to analyze the trade-offs involved in production inventory decisions and their impact on inventory holding costs, customer backorders, and service levels. The objective is to investigate the effect of system parameters such as product variety, production batch size on performance measures such as backorders and service levels.
The growing amount of multimedia content is making it hard for end users to find the relevant content. The goal of recommender systems is to assist the users by finding a small subset of relevant multimedia items for each user. State-of-the-art techniques for recommending content are very data-centric. The progress beyond the state-of-the-art presented in this book consists in introducing new parameters based on emotions and personality that explain a substantial part of variance in the end users'' preferences. The book covers the detection of emotions and personality factors of end users. The book then shows clearly how to use these user-centric data to model end users and thus improve the performance of a recommender system for images. The book will serve as a guideline and inspiration for practitioners and academics in content retrieval and affecting computing.
The work of this thesis falls in the field of the embedded real-time computer systems, more precisely the embedded software in the vehicles for the emergent driving assistance systems. The thesis work was supported by the European project (RoadSense) whose objective was the evaluation of the driving assistance systems (ADAS) for the next generations of vehicles. The three principal axes of the thesis are: First, the development of a dynamic reconfigurable middleware called SCOOT-R. Second, the development of end-to-end real-time scheduling strategies in order to schedule SCOOT-R operations with the main goal of end-to-end QoS guarantee. Finally, the development of feedback-based scheduling schemes to schedule driving assistance systems. The adaptation in this scheme is carried out according to the current driving situation and the driver behavior, which will further lead to the change of the associated driving assistance function''s criticalness.
The unique characteristics of multi-radio wireless mesh networks, such as shared nature of wireless channels, use of multiple radio interfaces, stationary mesh routers, and user-to-gateway traffic pattern, make them different from wired-networks and other wireless networks. In particular, the effects of an unbalanced load can cause rapid gateway overloading, center overloading, or channel overloading. Since most traffic in a WMN is destined towards gateways, traffic concentration at gateway nodes creates a load imbalance at certain gateways, which in turn results in gateway overloading. The issue of center overloading refers to the nodes located near the geographical center of the network becoming overloaded in comparison to other nodes in the network. Finally, channel overloading in a multi-radio wireless mesh network refers to certain channels becoming overloaded compared to other channels. Thus, load balancing is necessary to avoid hot spots and to increase network utilization, as bad routes can exist for a long time in a static network and result in congestion and inefficient use of network resources.
Military organizations, similarly to other organizations that are concerned about the security of information exchanges, with high energy efficiency have always heavily relied on secure exchanges of short messages. A reliable system for such message exchanges is considered to be a particular strength for such organizations. In this book two protocols that are used, first one is Secure Message Transmission (SMT) protocol in order to secure the data transmission phase by tailoring an end-to-end secure data forwarding protocol to the MANET communication requirements and increase the reliability through transmitting the messages in multiple paths with minimal redundancy. As a second one Concentration is on Energy Efficient topology in Ad-hoc networks. The topology that is used in this case is Demand Based Energy efficient Topology (DBET) to reduce the energy consumption for Mobile Ad-hoc network, by dynamically adjusting the topology for various network traffic conditions. Simulations were conducted using the NS2 network simulator. Nodes in the network were configured to use 802.11 radios with a bandwidth of 2 Mbps and a nominal range of 250 m.
This work is about a bold step towards bringing proven security to embedded systems. Most present day embedded systems are shipped into market without any strong security primitives. Those that do have security; if at all, the security is propriety and does not conform to standards. IPsec is a well proven suite of standards for computer communication, over TCP/IP. Why not use a reduced version of the IPsec standard, adopted intelligently for the embedded systems? The users of embedded systems would welcome such a protection, if the security rules are not violated and no weaknesses are introduced.
In this study some mathematical models in reliability systems were investigated, such as series , parallel , series – parallel , parallel – series and complex systems by using some methods such as ; path tracing method, reduction to series element method ,minimal cut method , event space method , composite method and decomposition method . A comparison was done between those methods to get the best reliability. Reliability allocation had been studied .The aim of it is to establish a goal or objective for the reliability of each component. Two methods had been used to find the reliability allocation for series and complex systems. This study also discuss the reliability for non – maintenance depending on the application of Markov and non – Markov models.
Three decades ago, many diagnosis knowledge-based systems have been developed in different diagnostic domains. The diagnosis systems may provide different solutions for the same problem. These solutions mirror different viewpoints for different experts. This book presents a framework for multi-agent diagnosis system as an effective solution for the problem of acquiring different diagnostic results when consulting different diagnostic systems. The framework defines how the agents can negotiate, and cooperate together to resolve the belief conflict that may arise among them. The framework can be applied in any diagnostic domain. Also, the book presents a system prototype in the plant disorders domain. The prototype has been developed on the basis of this framework as a real case study. This book is very interesting for post-graduate students and researchers who have interested in the research areas of multi-agent systems, conflict resolution, diagnosis systems, and agent interaction protocols. The book is also important for practitioners who are developing standalone knowledge-based systems and want to take step forward to developing multi-agent systems.