This paper presents control and coordination algorithms for groups of vehicles. The focus is on autonomous vehicle networks performing distributed sensing tasks where each vehicle plays the role of a mobile tunable sensor. The paper proposes gradient descent algorithms for a class of utility functions which encode optimal coverage and sensing policies. The resulting closed-loop behavior is adaptive, distributed, asynchronous, and verifiably correct.

Abstract. This paper presents snapshot algorithms for determining a consistent global state of a distributed system without significantly affecting the underlying computation. These algorithms do not require channels to be FIFO or messages to be acknowledged. Only a small amount of storage is needed. An important application of a snapshot algorithm is Global Virtual Time determination for distributed simulations. The paper proposes new and efficient Global Virtual Time approximation schemes based on snapshot algorithms and distributed termination detection principles. 1

We present a model of distributed computation which is based on a fragment of the pi-calculus relying on asynchronous communication. We enrich the model with the following features: the explicit distribution of processes to locations, the failure of locations and their detection, and the mobility of processes. Our contributions are two folds. At the specification level, we give a synthetic and flexible formalization of the features mentioned above. At the verification level, we provide original methods to reason about the bisimilarity of processes in the presence of failures.

Event-based systems are developed and used to integrate components in loosely coupled systems. Research and product development focused so far on e#ciency issues but neglected methodological support to build such systems. In this article, the modular design and implementation of an event system is presented which supports scopes and event mappings, two new and powerful structuring methods that facilitate engineering and coordination of components in event-based systems. We give a

The development of practical, localized algorithms is probably the most needed and most challenging task in wireless ad-hoc sensor networks (WASNs). Localized algorithms are a special type of distributed algorithms where only a subset of nodes in the WASN participate in sensing, communication, and computation. We have developed a generic localized algorithm for solving optimization problems in wireless ad-hoc networks that has five components: (i) data acquisition mechanism, (ii) optimization mechanism, (iii) search expansion rules, (iv) bounding conditions, and (v) termination rules. The main idea is to request and process data only locally and only from nodes who are likely to contribute to rapid formation of the final solution. The approach enables two types of optimization: The first, guarantees the fraction of nodes that are contacted while optimizing for solution quality. The second, provides guarantees on solution quality while minimizing the number of nodes that are contacted and/or amount of communication. This localized optimization approach is applied to two fundamental problems in sensor networks: location discovery and exposure-based coverage. We demonstrate its effectiveness on a number of examples.

Fault tolerance in distributed computing is a wide area with a significant body of literature that is vastly diverse in methodology and terminology. This paper aims at structuring the area and thus guiding readers into this interesting field. We use a formal approach to define important terms like fault, fault tolerance, and redundancy. This leads to four distinct forms of fault tolerance and to two main phases in achieving them: detection and correction. We show that this can help to reveal inherently fundamental structures that contribute to understanding and unifying methods and terminology. By doing this, we survey many existing methodologies and discuss their relations. The underlying system model is the close-to-reality asynchronous message-passing model of distributed computing.

Most existing work on sensor networks concentrates on finding efficient ways to forward data from the information source to the data centers, and not much work has been done on collecting local data and generating the data report. This paper studies this issue by proposing techniques to detect and track a mobile target. We introduce the concept of dynamic convoy tree-based collaboration, and formalize it as a multiple objective optimization problem which needs to find a convoy tree sequence with high tree coverage and low energy consumption. We propose an optimal solution which achieves 100% coverage and minimizes the energy consumption under certain ideal situations. Considering the real constraints of a sensor network, we propose several practical implementations: the conservative scheme and the prediction-based scheme for tree expansion and pruning; the sequential and the localized reconfiguration schemes for tree reconfiguration. Extensive experiments are conducted to compare the practical implementations and the optimal solution. The results show that the prediction-based scheme outperforms the conservative scheme and it can achieve similar coverage and energy consumption to the optimal solution. The experiments also show that the localized reconfiguration scheme outperforms the sequential reconfiguration scheme when the node density is high, and the trend is reversed when the node density is low.

Execution of MPI applications on clusters and Grid deployments suffering from node and network failures motivates the use of fault tolerant MPI implementations. We present MPICH-V2 (the second protocol of MPICH-V project), an automatic fault tolerant MPI implementation using an innovative protocol that removes the most limiting factor of the pessimistic message logging approach: reliable logging of in transit messages. MPICH-V2 relies on uncoordinated checkpointing, sender based message logging and remote reliable logging of message logical clocks. This paper presents the architecture of MPICH-V2, its theoretical foundation and the performance of the implementation. We compare MPICH-V2 to MPICH-V1 and MPICH-P4 evaluating a) its point-to-point performance, b) the performance for the NAS benchmarks, c) the application performance when many faults occur during the execution. Experimental results demonstrate that MPICH-V2 provides performance close to MPICH-P4 for applications using large messages while reducing dramatically the number of reliable nodes compared to MPICH-V1. 1

) Alain Bui, 1 Ajoy K. Datta, 2 Franck Petit, 1 Vincent Villain 1 1 LaRIA, Universit e de Picardie Jules Verne, France 2 Department of Computer Science, University of Nevada, Las Vegas Abstract In this paper, we introduce the notion of snapstabilization. A snap-stabilizing algorithm protocol guarantees that, starting from an arbitrary system configuration, the protocol always behaves according to its specification. So, a snap-stabilizing protocol is a self-stabilizing protocol which stabilizes in 0 steps. We propose a snap-stabilizing Propagation of Information with Feedback (PIF) scheme on a rooted tree network. We call this scheme Propagation of information with Feedback and Cleaning (PFC). We present two algorithms. The first one is a basic PFC scheme which is inherently snapstabilizing. However, it can be delayed O(h 2 ) steps (where h is the height of the tree) due to some undesirable local states. The second algorithm improves the worst delay of the basic PFC algori...

We investigate the semantics of messages, and argue that the meaning of a message is naturally and usefully given in terms of how it a#ects the knowledge of the agents involved in the communication. We see that the semantics depends on the protocol used by the agents, and this leads us to knowledge based specification of protocols. While these notions are natural for distributed computations, we suggest that the considerations discussed here may be relevant in more general linguistic contexts.