Catálogo de publicaciones - libros

As the resources in a wireless network are limited and freely shared by all network users, Call Admission Control (CAC) plays a significant role in providing the Quality of Service (QoS) in wireless networks. However, when the network is congested with too many users, traditional CAC that mainly focuses on the tradeoff between new call blocking probability and handoff call dropping probability cannot guarantee QoS requirements to users. To address this issue, this paper proposes a dual level congestion control scheme which considers not only the call level admission control but also the user’s decision to enter the network or not during the network traffic burst interval (we call it as user-level burst control). We adopt the economical terms of externality and introduce a total user utility function to formally model the user’s perceived QoS metric. Our simulation shows that the weighted blocking probability (Pb) of our scheme can decreases 70~80% than traditional CAC systems and increase the total user utility to 2~3 times.

Pervasive computing is widely researched and a large number of context-aware applications have been built in the recent years. However, correctness of contexts and fault handling of these applications have always been ignored. This paper proposes a transaction model for context-aware applications. In this model, context-aware applications are organized as a number of logic units and each unit may have a compensation module, which will be executed when errors or exceptions occur in context-aware applications in order to minimize the bad infection. This model supports nested scopes and the number of levels of subtransactions is unlimited. We also present an implementation of this transaction model, which is specialized for context-aware use.

Recently, there is an increasing need for researchers in engineering to share the result of the experiment without having to visit the experiment facilities. Especially in the civil engineering, researchers feel the need for participating in a number of experiments conducted at distant places. In addition, it has been suggested that high-cost facilities should be used by remote researchers for the high utilization rate. This paper proposes a remote experiment environment in civil engineering that are being developed in a project called Korea Construction Engineering Development(KOCED), which connects major civil engineering experiment facilities using grid technology. This environment enables researchers to participate in a remote experiment, and allows the experiment results shared by remote researchers automatically. Then, based on the suggested environment, we designed a hybrid test facility that involves two physical experiment facility sites and one numerical simulation site that are geographically apart. Then, we implemented its prototype and ran some tests, which showed a possibility of grid-based civil engineering experiment.

Ad hoc network is an infra structure less network, which is formed by mobile devices like laptops, PDAs, cell phones etc. Each device has different computational capability, power, hardware and software, which forms a heterogeneous network. These devices can be integrated to form an infrastructure known as grid. A grid integrates and coordinates resources and users that are within the same network with different capabilities. Hence we can visualize a grid over an ad hoc network that effectively utilizes the heterogeneity in the mobile devices. The major challenge in forming a grid over an ad hoc network is the mobility of the nodes. In this paper, we address the challenges due to mobility by considering a trace model for the movement of the nodes. Next, we demonstrate the feasibility of forming a grid over a mobile ad hoc network by proposing lightweight algorithms for grid formation, resource discovery, negotiation, job scheduling, and resource sharing. We have analyzed the performance of mobile ad hoc grid both by using a theoretical model and by simulation. The results point to a promising approach to form a mobile ad hoc grid.

As grid infrastructures become more dynamic in order to cope with the uncertainty of demand, they are becoming extremely difficult to manage. At the Belfast e-Science Centre, we are attempting to address this issue by developing Self Managing Grid Middleware. This paper gives an overview of the middleware and focuses on the design, implementation and evaluation of a Resource Manager. Also in this paper we will see how our approach, which is based on federated UDDI registries, has enabled us to implement some of the desired features of next generation grid software.

During the past several years, the grid application executed same jobs on one or more hosts in parallel, but the recent grid application is requested to execute different jobs linearly. That is, the grid application takes the form of workflow application. In general, efficient scheduling of workflow applications is based on heuristic scheduling method. The heuristic considering relation between hosts would improve execution time in workflow applications. But because of the heterogeneity and dynamic nature of grid resources, it is hard to predict the performance of grid application. In addition, it is necessary to deal with user’s QoS as like performance guarantee. In this paper, we propose a service model for predicting performance and an adaptive workflow scheduling strategy, which uses maximum flow algorithms for the relation of services and user’s QoS. Experimental results show that the performance of our proposed scheduling strategy is better than common-used greedy strategies.

It is important to debug unintended data races in OpenMP programs efficiently, because such programs are often complex and long-running. Previous tools for detecting the races does not provide any effective facility for understanding the complexity of threads involved in the reported races. This paper presents a thread visualization tool to present a partial order of threads executed in the traced programs with a scalable graph of abstract threads upon a three-dimensional cone. The scalable thread visualization is proved to be effective in debugging races using a set of synthetic programs.

Message races, which can cause nondeterministic executions of a parallel program, should be detected for debugging because nondeterminism makes debugging parallel programs a difficult task. Even though there are some tools to detect message races in MPI programs, they do not provide practical information to locate and debug message races in MPI programs. In this paper, we present an on-the-fly detection tool, which is MPIRace-Check, for debugging MPI programs written in C language. MPIRace-Check detects and reports all race conditions in all processes by checking the concurrency of the communication between processes. Also it reports the message races with some practical information such as the line number of a source code, the processes number, and the channel information which are involved in the races. By providing those information, it lets programmers distinguish of unintended races among the reported races, and lets the programmers know directly where the races occur in a huge source code. In the experiment we will show that MPIRace-Check detects the races using some testing programs as well as the tool is efficient.

Position-based routing has been proven to be a scalable and efficient solution for packet routing in mobile ad hoc networks (MANETs) by utilizing location information of mobile nodes. The location service provides geographic locations for all nodes and is therefore critical to position-based routing. In general, the control overhead in a position-based routing protocol is mainly dominated by location updates. In this paper, we propose a location service called Modified Grid Location Service (MGLS), which employs a binary grid partitioning scheme to reduce the control overhead associated with the location management and supports large scale ad hoc networks. We then use a theoretical model to analyze both MGLS and GLS. Both theoretical analysis and simulation results show that MGLS can reduce the location update overhead in location services.

Pervasive Grids emerge as a new paradigm for providing nomadic users with ubiquitous access to digital information and computing resources. However, pervasive grids arise a number of crucial issues related to privacy and security, especially authentication and access control, which constitute the security front-end.

In this paper, we propose a trust based model of authentication and access control that allows nomadic users to roam from site to site and to gain access to surrounding/remote resources wrt her status in her home site and to the local policy of the site where she is standing. This model is supported by a software architecture called Chameleon.

The Chameleon permits users to access grid resources and to implement adhoc interactions with the local grid site.