Showing papers by "Nitin H. Vaidya published in 2012"

Iterative approximate byzantine consensus in arbitrary directed graphs

Abstract: This paper proves a necessary and sufficient condition for the existence of iterative, algorithms that achieve approximate Byzantine consensus in arbitrary directed graphs, where each directed edge represents a communication channel between a pair of nodes. The class of iterative algorithms considered in this paper ensures that, after each iteration of the algorithm, the state of each fault-free node remains in the convex hull of the states of the fault-free nodes at the end of the previous iteration. The following convergence requirement is imposed: for any e > 0, after a sufficiently large number of iterations, the states of the fault-free nodes are guaranteed to be within e of each other. To the best of our knowledge, tight necessary and sufficient conditions for the existence of such iterative consensus algorithms in synchronous arbitrary point-to-point networks in presence of Byzantine faults, have not been developed previously. The methodology and results presented in this paper can also be extended to asynchronous systems.

Resilient Networked Control of Distributed Energy Resources

Abstract: This paper considers networked systems and develops distributed algorithm that is resilient against potential packet drops in the communication links between system components. We apply this algorithm to the problem of coordinating distributed energy resources (DERs) for the provision of ancillary services in electrical networks, e.g., reactive power support for voltage control. In this problem, each system component can contribute a certain amount of active and/or reactive power, bounded from above and (possibly) below by capacity constraints, and the objective is to coordinate the components so as to collectively provide a predetermined total amount of active and/or reactive power. In the algorithm we propose to address this problem, each DER maintains a set of variables and updates them through information exchange with neighboring DERs. We show that, as long as the underlying graph that describes the information exchange between components is strongly connected, and the predetermined total amount of active and/or reactive power does not violate (upper or lower) total capacity constraints, DERs can use this approach to calculate, in a distributed fashion, their fair contribution (subject to their capacity constraints). We show that the proposed algorithms reach almost surely convergence to the fair solution, even in the presence of communication link failures.

Matrix Representation of Iterative Approximate Byzantine Consensus in Directed Graphs

Abstract: : This paper presents a proof of correctness of an iterative approximate Byzantine consensus (IABC) algorithm for directed graphs. The iterative algorithm allows fault- free nodes to reach approximate consensus despite the presence of up to f Byzantine faults. Necessary conditions on the underlying network graph for the existence of a correct IABC algorithm were shown in our recent work [15, 16]. [15] also analyzed a specific IABC algorithm and showed that it performs correctly in any network graph that satisfies the necessary condition, proving that the necessary condition is also sufficient. In this paper, we present an alternate proof of correctness of the IABC algorithm using a familiar technique based on transition matrices [9, 3, 17, 19]. The key contribution of this paper is to exploit the following observation: for a given evolution of the state vector corresponding to the state of the fault-free nodes many alternate state transition matrices may be chosen to model that evolution correctly. For a given state evolution, we identify one approach to suitably design the transition matrices so that the standard tools for proving convergence can be applied to the Byzantine fault-tolerant algorithm as well. In particular, the transition matrix for each iteration is designed such that each row of the matrix contains a large enough number of elements that are bounded away from 0.

Iterative Approximate Byzantine Consensus under a Generalized Fault Model

Abstract: In this work, we consider a generalized fault model [7,9,5] that can be used to represent a wide range of failure scenarios, including correlated failures and non-uniform node reliabilities. Under the generalized fault model, we explore iterative approximate Byzantine consensus (IABC) algorithms [15] in arbitrary directed networks. We prove a tight necessary and sufficient condition on the underlying communication graph for the existence of IABC algorithms.

Iterative Approximate Byzantine Consensus under a Generalized Fault Model

Abstract: In this work, we consider a generalized fault model that can be used to represent a wide range of failure scenarios, including correlated failures and non-uniform node reliabilities. This fault model is general in the sense that fault models studied in prior related work, such as f -total and f -local models, are special cases of the generalized fault model. Under the generalized fault model, we explore iterative approximate Byzantine consensus (IABC) algorithms in arbitrary directed networks. We prove a necessary and sufficient condition for the existence of IABC algorithms. The use of the generalized fault model helps to gain a better understanding of IABC algorithms.

RFID reader collision problem: performance analysis and medium access

Abstract: The RFID reader collision problem, in which an RFID reader's interrogation is interfered by other concurrent readers' transmission, is considered an important issue to reliable operation and thus to the wide-spread deployment of RFID networks. In this paper, we present modeling and analysis of the RFID reader collision problem. We observe asymmetry between an RFID reader's and a tag's communication capabilities and develop an RFID radio model based on the asymmetry. By the model, we characterize the spatial reuse of RFID reader networks, and derive concurrent interrogation distance beyond which readers can transmit simultaneously without causing collision and the carrier sense threshold corresponding to the distance. We also examine the dual-channel mode where available bandwidth is divided into two channels by which reader-to-tag communication and tag-to-reader communication are separated. We analyze and evaluate the performance of the dual-channel mode in terms of spatial reuse and interrogation completion time. Copyright © 2010 John Wiley & Sons, Ltd.

Iterative Approximate Byzantine Consensus in Arbitrary Directed Graphs - Part II: Synchronous and Asynchronous Systems

Abstract: This report contains two related sets of results with different assumptions on synchrony. The first part is about iterative algorithms in synchronous systems. Following our previous work on synchronous iterative approximate Byzantine consensus (IABC) algorithms, we provide a more intuitive tight necessary and sufficient condition for the existence of such algorithms in synchronous networks1. We believe this condition and the previous results also hold in partially asynchronous algorithmic model. In the second part of the report, we explore the problem in asynchronous networks. While the traditional Byzantine consensus is not solvable in asynchronous systems, approximate Byzantine consensus can be solved using iterative algorithms.

Robust average consensus over packet dropping links: Analysis via coefficients of ergodicity

Abstract: We consider a networked system in which each component (node) iteratively exchanges information with its neighbors according to an arbitrary, possibly directed topology. Based on an iterative exchange of (local and possibly directed) information, we develop an average-consensus distributed algorithm that is robust to unreliable (packet-dropping) communication links. By introducing virtual nodes, we show that the execution of the proposed algorithm is mathematically equivalent to a finite inhomogenous Markov chain. Then, by using coefficients of ergodicity, we can prove convergence of the robust distributed algorithm to the exact average, in the presence of packet drops and under a very broad set of conditions.

Experimental performance comparison of Byzantine Fault-Tolerant protocols for data centers

Abstract: In this paper, we compare performance of several Byzantine agreement algorithms, including NCBA, a network coding based algorithm. Unlike existing practical BFT protocols such as PBFT by Castro and Liskov [1], which utilize collision-resistant hash functions to reduce traffic load for BFT, NCBA uses a computationally efficient error-detection network coding scheme. Since NCBA does not rely on any hash function, it is always correct rather than correct only with high probability as PBFT. Through extensive experiments, we verified that NCBA performs at least as well as Digest, without relying on any cryptographic assumption on the hardness of breaking the hash function. To the best of our knowledge, this is the first implementation of BFT with network coding.

Workload-aware opportunistic routing in multi-channel, multi-radio wireless mesh networks

Abstract: Opportunistic routing emerged as a novel technique to cope with the problem of highly unpredictable and lossy wireless channels in urban wireless mesh networks. However, existing opportunistic routing protocols only consider single-radio wireless nodes, and assume that all the nodes work on the same channel, without exploiting possible concurrent transmissions by multi-radio nodes over orthogonal channels provided by IEEE 802.11 protocols. Examples show that simply integrating existing channel assignment schemes and the opportunistic routing technique may not achieve satisfactory system performance. In this paper, we present WACA, which is a Workload-Aware Channel Assignment algorithm for opportunistic routing in multi-channel, multi-radio wireless mesh networks. Evaluation results show that WACA always achieves highest average throughput among the evaluated algorithms, and its median throughput is at least 16.1% higher than the compared ones.

Resilient average consensus in the presence of heterogeneous packet dropping links

Abstract: We address the average-consensus problem for a distributed system whose components (nodes) can exchange information via unreliable interconnections (edges) that form an arbitrary, possibly directed topology (digraph). We consider a general setting where heterogeneous communication links may drop packets with generally unequal probabilities, independently between different links. We develop a distributed linear-iterative algorithm in which nodes maintain and update certain values based on the corresponding values they successfully receive from their in-neighbors. We demonstrate that, even when communication links drop packets with unequal probabilities, the proposed algorithm allows nodes to asymptotically reach average-consensus almost surely, as long as the underlying (possibly directed) communication topology forms a strongly connected digraph. Additionally, we provide a bound on the algorithm convergence rate.

Exact Byzantine Consensus in Directed Graphs

Abstract: For synchronous point-to-point n-node networks of undirected links, it has been previously shown that, to achieve consensus in presence of up to f Byzantine faults, the following two conditions are together necessary and sufficient: (i) n ≥ 3f + 1 and (ii) network connectivity greater than 2f . The first condition, that is, n ≥ 3f + 1, is known to be necessary for directed graphs as well. On the other hand, the second condition on connectivity is not necessary for directed graphs. So far, tight necessary and sufficient condition for Byzantine consensus in directed graphs has not been developed. This paper presents tight necessary and sufficient condition for achieving Byzantine consensus in synchronous networks that can be represented as directed graphs. We provide a constructive proof of sufficiency by presenting a new Byzantine consensus algorithm for directed graphs. Further work is needed to improve the message overhead of Byzantine consensus in directed graphs.

A new ‘Direction’ for source location privacy in wireless sensor networks'

Abstract: Preserving source location privacy in wireless sensor networks can be critical for several practical applications. Existing solutions proposed specifically for sensor networks rely on a combination of dynamic routing and dummy traffic to hide real event messages. While some privacy protection guarantees can be given, these solutions also tend to be expensive due to fake transmissions and non-shortest path routing overheads. In this paper, we propose a novel idea, of using a combination of directional antennas, transmit power control and information compression to provide lightweight and energy-efficient source location privacy. We discuss the adversary model extensively and then carefully layout characteristics of a realistic adversary. We show how use of directional antennas makes eavesdropping more costly for a realistic adversary and establish relationships between probability of compromise of location privacy, characteristics of directional antennas and size of the adversary's eavesdropping network. Finally, we show how a simple information compression measure can greatly reduce message latency and prolong battery life by conserving energy. Results of extensive simulations in NS2, with our realistic directional antenna add-on, show that compared to existing solutions, we can achieve comparable privacy protection, better message latency, delivery ratio and many orders of magnitude improvement in energy consumption.

Exploiting opportunistic overhearing to improve performance of mutual exclusion in wireless ad hoc networks

Abstract: We design two mutual exclusion algorithms for wireless networks Our mutual exclusion algorithms are distributed token based algorithms which exploit the opportunistic message overhearing in wireless networks One of the algorithms is based on overhearing of token transmission In the other algorithm, overhearing of both token and request messages is exploited The design goal is to decrease the number of transmitted messages and delay per critical section entry using the information obtained from overheard messages

A link layer protocol and link-state routing protocol suite for multi-channel ad hoc networks

Abstract: We propose a link layer protocol and link-state routing protocol suite for multi-channel ad hoc networks. The proposed protocol suite addresses several practical issues that arise when nodes equipped with two radio interfaces want to utilize available channels. The routing layer makes a hybrid channel assignment where one interface is fixed and the other is switchable. Based on that, the routing layer runs a shortest path routing algorithm augmented with channel diversity. The link layer implements a slotted structure to minimize broadcast overhead inherent to multi-channel networks. By using flow- and packet-level simulators, we make some important observations. First, a hybrid channel assignment is good for connectivity and amenable to shortest-path routing. Second, seeking a shortest-path can be a better routing strategy in terms of global system throughput than complex channel-diverse routing. Third, channel switching delay is not a throttling factor in terms of global system throughput. Copyright © 2010 John Wiley & Sons, Ltd.

Iterative Approximate Byzantine Consensus in Arbitrary Directed Graphs

Abstract: In this paper, we explore the problem of iterative approximate Byzantine consensus in arbitrary directed graphs. In particular, we prove a necessary and sufficient condition for the existence of iterative byzantine consensus algorithms. Additionally, we use our sufficient condition to examine whether such algorithms exist for some specific graphs.

Broadcast Using Certified Propagation Algorithm in Presence of Byzantine Faults

Abstract: We explore the correctness of the Certified Propagation Algorithm (CPA) [6, 1, 8, 5] in solving broadcast with locally bounded Byzantine faults. CPA allows the nodes to use only local information regarding the network topology. We provide a tight necessary and sufficient condition on the network topology for the correctness of CPA. To the best of our knowledge, this work is the first to solve the open problem in [8]. We also present some simple extensions of this result

Byzantine broadcast in point-to-point networks using local linear coding

Abstract: The goal of Byzantine Broadcast (BB) is to allow a set of fault-free nodes to agree on information that a source node wants to broadcast to them, in the presence of Byzantine faulty nodes. We consider design of efficient algorithms for BB in synchronous point-to-point networks, where the rate of transmission over each communication link is limited by its "link capacity". The throughput of a particular BB algorithm is defined as the average number of bits that can be reliably broadcast to all fault-free nodes per unit time using the algorithm without violating the link capacity constraints. The capacity of BB in a given network is then defined as the supremum of all achievable BB throughputs in the given network, over all possible BB algorithms. We develop NAB - a Network-Aware BB algorithm - for tolerating f faults in arbitrary point-to-point networks consisting of f ≥ 3f+1 nodes and having ≥ 2f+1 directed node disjoint paths from each node i to each node j. We also prove an upper bound on the capacity of BB, and conclude that NAB can achieve throughput at least 1/3 of the capacity. When the network satisfies an additional condition, NAB can achieve throughput at least 1/2 of the capacity. To the best of our knowledge, NAB is the first algorithm that can achieve a constant fraction of capacity of Byzantine Broadcast (BB) in general point-to-point networks.

Watchdogs to the rescue: Securing wireless TCP

Abstract: In this paper, we make a case for using watchdogs to protect against misbehavior in dense wireless networks. We introduce “Generalized Watchdogs” and identify when and how watchdogs can be necessary and sufficient against misbehavior. We study feasibility of watchdog approach and show that the order of capacity bounds is preserved asymptotically even with watchdogs. We use generalized watchdogs to design protocols to improve both security and performance of TCP over wireless networks such that the application at the destination never accepts a corrupted packet and we achieve this without modifying TCP. We show that a strict dependence on availability and success of watchdogs can lead to “watchdog induced losses” and establish their effects on TCP throughput. We then propose solutions to deal with these losses and make watchdogs intelligent so they can tune the overheads incurred. With hop-by-hop verification of packet correctness, we ensure that tampered packets are not forwarded in the network and thus save potential wastage of network resources. We use NS-2 simulations of both controlled as well as realistic network scenarios, to show that watchdogs can provide simple, lightweight and reliable means of misbehavior detection, tolerance and most importantly “deterrence” while saving costs of security infrastructure. With a combination of intelligent watchdogs and source coding, and by leveraging route adaptation, our scheme achieves twice the throughput of a cryptographic alternative and that too in presence of as high as 30% packet tampering.

Token-DCF: An Opportunistic MAC protocol for Wireless Networks

Abstract: IEEE 802.11 DCF is the MAC protocol currently used in wireless LANs. 802.11 DCF is inefficient due to two types of overhead; channel idle time and collision time. This paper presents the design and performance evaluation of an efficient MAC protocol for wireless networks, called Token-DCF. Token-DCF decreases both idle time and collision time. In Token-DCF, each station keeps track of neighboring links' queue length by overhearing of transmitted packets on the wireless medium. The result is then used to assign privileges to the network stations. A privileged station does not follow the backoff mechanism and transmits immediately after the channel is sensed idle. Our simulation results show that Token-DCF can significantly improve channel utilization, system throughput and channel access delay over 802.11 DCF.

Capacity of Byzantine consensus in capacity limited point-to-point networks

Abstract: In this paper, we investigate the problem of maximizing the throughput, i.e., achieving capacity, of Byzantine consensus in point-to-point networks, in which each link has a capacity constraint. We derive an upper bound of the capacity of consensus in general point-to-point networks, and prove its tightness in 4-node complete networks by construction. We also provide a probabilistically correct algorithm that achieves the upper bound in general networks.

Parameter-independent Iterative Approximate Byzantine Consensus

Abstract: : In this work, we explore iterative approximate Byzantine consensus algorithms that do not make explicit use of the global parameter of the graph, i.e., the upper-bound on the number of faults, f .

Recoverable Distributed Shared Memory Using the Competitive Update Protocol

Abstract: In this paper, we propose a recoverable DSM that uses a competitive update protocol. In this update protocol, multiple copies of each page may be maintainedat different nodes. However, it is also possible fora page to exist in only one node, as some copies of the page may be invalidated. We propose an implementation that makes the competitive update protocol recoverable from a single node failure, by guaranteeing that at least two copies of each page exist. The paper presents preliminary evaluation of the recoverable DSM (using simulation). It is shown that the message overhead of making the DSM recoverable is small.