Catálogo de publicaciones - libros

We present , a replication method for unstructured Peer-to-Peer overlays. The goal of our method is to achieve real-time replication of even the most sparsely located content relative to demand. adaptively or the replica set of an object in order to improve the sharing process and achieve a low load distribution among the providers. To achieve that, it utilizes search knowledge to identify possible replication targets inside query-intensive areas of the overlay. We present detailed simulation results where exhibits both efficiency and robustness over the number of requesters and the respective request rates. The scheme proves particularly useful in the event of flash crowds, managing to quickly adapt to sudden surges in load.

Peer-to-Peer data structures (P2P data structures) let a large number of anonymous peers share the data-management workload. A common assumption behind such systems is that peers behave cooperatively. But as with many distributed systems where participation is voluntary, and the participants are not clearly observable, unreliable behavior is the dominant strategy. This calls for reputation systems that help peers choose reliable peers to interact with. However, if peers exchange feedback on experiences with other peers, spoof feedback becomes possible, compromising the reputation system. In this paper we propose and evaluate measures against spoof feedback in P2P data structures. While others have investigated mechanisms for truthtelling recently, we are not aware of any studies in P2P environments. The problem is more difficult in our context because detecting unreliable peers is more difficult as well. On the other hand, a peer can observe the utility of feedback obtained from other peers, and our approach takes advantage of this. To assess the effectiveness of our approach, we have conducted extensive analytical and experimental evaluations. As a result, truthful feedback tends to have a much higher weight than spoof feedback, and collaboration attacks are difficult to carry out under our approach.

In this paper, we will present an efficient approach for distributed inference. We use belief propagation’s message-passing algorithm on top of a DHT storing a Bayesian network. Nodes in the DHT run a variant of the spring relaxation algorithm to redistribute the Bayesian network among them. Thereafter correlated data is stored close to each other reducing the message cost for inference. We simulated our approach in Matlab and show the message reduction and the achieved load balance for random, tree-shaped, and scale-free Bayesian networks of different sizes.

As possible application, we envision a distributed software knowledge base maintaining encountered software bugs under users’ system configurations together with possible solutions for other users having similar problems. Users would not only be able to repair their system but also to foresee possible problems if they would install software updates or new applications.

At the early stages of the cooperative information system development one of the major problems is to explore the space of alternative ways of assignment and delegations of goals among system actors. The exploration process should be guided by a number of criteria to determine whether the adopted alternative is good-enough. This paper frames the problem of designing actor dependency networks as a multi-agent planning problem and adopts an off-the-shelf planner to offer a tool (P-Tool) that generates alternative actor dependency networks, and evaluates them in terms of metrics derived from Game Theory literature. As well, we offer preliminary experimental results on the scalability of the approach.

Current technologies are continuously evolving and software companies need to adapt their processes to these changes. Such adaptation often requires new investments in training and development. To address this issue, OMG defined a model driven development approach (MDD) which insulates business and application logic from technology evolution. Current MDD approaches falls short in fully derive implementation from models described at a high abstraction level. We propose a controlled natural language to complement UML models as an action specification language. In this article, we describe the language, its impact on systems development and the tools developed to support it. In order to demonstrate the language usability we present an application example.

Shared mutable information objects called u-forms provide an attractive foundation on which to build collaborative systems. As we scale up such systems from small fully-connected workgroups to large, highly distributed, and partially disconnected groups, we have found that peer-to-peer technology and optimistic replication strategies provide a cost-effective mechanism for maintaining good performance. Unfortunately, such systems present well-known coordination and consistency problems. This paper discusses strategies for addressing those difficulties at different levels of the system design, focusing on providing solutions in the information architecture rather than at the infrastructure layer. Addressing problems at this higher layer allows greater freedom in design, and simplifies moving from one infrastructural base to another as technology evolves. Our primary strategy is to enable robust decentralized and asynchronous collaboration while designing architectures that do not rely on two users writing to the same u-form at the same time in different venues. Techniques are provided for simple messaging, collaborative maintenance of collections, indexing supporting rich query, and stand-off annotation and elaboration of third-party datasets. We outline the application of these techniques in a working collaborative system.

Despite computers’ widespread use for personal applications, very few programming frameworks exist for creating synchronous collaborative applications. Existing research in CSCW (computer supported cooperative work), specifically approaches that attempt to make current application implementations collaboration-aware, are difficult to implement for two reasons: the systems are focused too narrowly (e.g., on Internet-only applications), or the systems are simply too complicated to be adopted (e.g., they are hard to set up and adapt to concrete applications). Enabling real-time collaboration demands lightweight, modular middleware——that enables the fine-grained interactions required by collaborative applications. In this paper, we introduce sliverware and give a specific example in the guise of a that multiplexes input from several users into a single stream that each user receives just like input from a normal keyboard. The result is simple, real-time collaboration based on a shared, distributed view of data that enables rapid development of highly coupled coordinating applications.

Groupware applications allow a distributed group of human users to work apart together over a computer network. They are difficult to develop due to the needs to suit a range of collaboration tasks that are often with diverse and evolutionary requirements. To address this problem, we propose a new framework in which shared data components conforming to a well-defined interface can be dynamically plugged in for flexible sharing, and a simple transformation tool is provided such that the myriad of industrial collaboration-transparent components can be transformed into shared components. The validity of our framework is evaluated by building a suite of typical collaboration tools such as group editors. Under our framework, most components in the Java Development Kit (JDK) can be transformed automatically for prototyping collaboration tools. With minimal manual work, those tools can be adapted to achieve advanced flexibility, e.g., data and control components can be bound dynamically to switch control protocols.

In practice dependency management often suffers from labor intensity and complexity in creating and maintaining the dependency relations. Our method targets projects, where developers are geographically distributed and a wide range of tools is used. A conceptual domain model is used to inter-relate the development objects and to automate dependency link discovery. The proposed method is based on association of development objects with concepts from domain model. These associations are used to compute dependency among development objects, and are stepwise refined to direct dependency links.

A preliminary empirical evaluation of the method is conducted. The method is evaluated both on performance and psychological variables. The evaluation has been performed in laboratory settings using two real cases. The results, although preliminary, provide positive evidence about the ability of our method to automate discovery of dependency relations, the analysis indicates that the method is perceived to be easy to use and useful by its potential users.

Personalized recommendations in a mobile tourist information system suffer from a number of limitations. Most pronounced is the amount of initial user information needed to build a user model. In this paper, we adopt and extend the basic concepts of recommendation paradigms by exploiting a user’s personal information (e.g., preferences, travel histories) to replace the missing information. The designed algorithms are embedded as recommendation services in our TIP prototype. We report on the results of our analysis regarding effectiveness and performance of the recommendation algorithms. We show how a number of limiting factors were successfully eliminated by our new recommender strategies.