Gurdip Singh

Leader election in the presence of link failures

We study the problem of leader election in the presence of intermittent link failures. We assume that up to N/2-1 links incident on each node may fail during the execution of the protocol. We present a message optimal algorithm with message complexity O(N/sup 2/).

Model-checking Middleware-based Event-driven Real-time Embedded Software ?

Component frameworks such as the CORBA Component Model (CCM) and middleware services such as the CORBA Event Service are increasingly being used to build safety / mission-critical distributed real-time embedded (DRE) systems. In this paper, we present a novel model-checking infrastructure for checking global temporal properties of DRE systems built on top of a Real-Time CORBA Event Service using CCM architectures. We describe how (a) building support for OO structures and communication layers directly in an extensible model-checker and (b) leveraging domain properties related to priorities, scheduling, and timing can dramatically reduce the costs of checking realistic systems.

Leader election in complete networks

This paper presents protocols for leader election in complete networks. The protocols are message optimal and their time complexities are a significant improvement over currently known protocols for this problem. For asynchronous complete networks with sense of direction, we propose a protocol which requires <italic>O(N)</italic> messages and <italic>O(log N)</italic> time. For asynchronous complete network without sense of direction, we show that Ω(<italic>N/log N</italic>) is a lower bound on the time complexity of any message optimal election protocol and we present a family of protocols which requires <italic>O(Nk)</italic> messages and <italic>O(N/k)</italic> time, <italic>log N ≤ k ≤ N</italic>. Our results also improve the time complexity of several other related problems such as spanning tree construction, computing a global function, etc.

A Communication Protocol for a Vehicle Collision Warning System

Advances in wireless communication technologies have enabled vehicles on a highway to communicate in order to share state information and provide drivers with potential collision warnings. This paper proposes a protocol to support such a vehicle collision warning system. By sharing vehicle state information, the protocol is able to predict potential collisions and deliver warning messages to address different emergency scenarios. The protocol is based on the concept of Shortest Safety Distance between vehicles and the Safety Invariant. The protocol provides mechanisms to evaluate potential violations of the safety invariant and propagate warning messages to avoid them in a timely manner. It ensures that warning messages are propagated to all endangered vehicles in an emergency scenario in an efficient manner by keeping the number of messages low. We have simulated the system using the SPIN model checker to show the correctness of the protocol and its effectiveness in eliminating redundant messages.

A highly asynchronous minimum spanning tree protocol

SummaryIn this paper, we present an efficient distributed protocol for constructing a minimum-weight spanning tree (MST). Gallager, Humblet and Spira [5] proposed a protocol for this problem with time and message complexitiesO(N logN) andO(E+NlogN) respectively. A protocol withO(N) time complexity was proposed by Awerbuch [1]. We show that the time complexity of the protocol in [5] can also be expressed asO((D+d) logN), whereD is the maximum degree of a node andd is a diameter of the MST and therefore this protocol performs better than the protocol in [1] wheneverD+d<N/logN. We give a protocol which requiresO(min(N, (D+d)logN)) time andO(E+NlogNlogN/loglogN) messages. The protocol constructs a minimum spanning tree by growing disjoint subtrees of the MST (which are referred to asfragments). Fragments having the same minimum-weight outgoing edge are combined until a single fragment which spans the entire network remains. The protocols in [5] and [1] enforce a balanced growth of fragments. We relax the requirement of balanced growth and obtain a highly asynchronous protocol. In this protocol, fast growing fragments combine more often and there-fore speed up the execution.

Efficient leader election using sense of direction

Summary. This paper presents a protocol for leader election in complete networks with a sense of direction. Sense of direction provides nodes the capability of distinguishing between their incident links according to a global scheme. We propose a protocol for leader election which requires O(N) messages and O(log N) time. The protocol is message optimal and the time complexity is a significant improvement over currently known protocols for this problem.

Efficient distributed algorithms for leader election in complete networks

An efficient protocol for leader election in an asynchronous complete network is presented. The time complexity of the protocol is better than the currently known protocols for this problem. A message optimal protocol is presented that requires O(N/log N) time, where N is the number of nodes in the network. Also given is family of protocols with message and time complexities O(Nk) and O(N/k) respectively, where log N<or=k<or=N. Many problems such as spanning tree construction, computing a global function, etc., are equivalent to leader election in terms of their message and time complexities, and therefore the authors results also improve the time complexity of these problems.<<ETX>>

Event ordering in pervasive sensor networks

In a pervasive sensor system, application entities may use the available infrastructure resources such as sensing devices, actuators, wireless modules and gateways to accomplish various tasks. For example, a group of humans/robots may use the infrastructure to track the current location of the group members. Many such tasks require a service which keeps track of the order of occurrence of events in the application. Existing algorithms for this problem have been designed for traditional distributed systems using the layering paradigm where the algorithm is able to observe and intercept application messages. Pervasive applications, however, may not be structured as strictly layered systems. In this paper, we study the problem of event ordering in pervasive systems, and propose algorithms for detecting event ordering. Our first algorithm treats the entire network as a single graph, whereas the second algorithm operates in a hierarchical manner by subdividing the network into smaller groups. We present experimental evaluation of our algorithms using both simulation and deployment on a sensor testbed.

Leader Election in Complete Networks

Leader election is a fundamental problem in distributed computing and has a number of applications. This paper studies the problem of leader election in complete asynchronous networks. We present a message-optimal protocol that requires

A compositional approach for designing protocols

O(N\log{N})