arXiv subjects

Arnaud Casteigts

Publications and source records attributed to Arnaud Casteigts.

36 records · Page 2Linked to original sources

arXiv2019

Robustness: a New Form of Heredity Motivated by Dynamic Networks

We investigate a special case of hereditary property in graphs, referred to as {\em robustness}. A property (or structure) is called robust in a graph $G$ if it is inherited by all the connected spanning subgraphs of $G$. We motivate this definition using two different settings of dynamic networks. The first corresponds to networks of low dynamicity, where some links may be permanently removed so long as the network remains connected. The second corresponds to highly-dynamic networks, where communication links appear and disappear arbitrarily often, subject only to the requirement that the entities are temporally connected in a recurrent fashion ({\it i.e.} they can always reach each other through temporal paths). Each context induces a different interpretation of the notion of robustness. We start by motivating the definition and discussing the two interpretations, after what we consider the notion independently from its interpretation, taking as our focus the robustness of {\em maximal independent sets} (MIS). A graph may or may not admit a robust MIS. We characterize the set of graphs \forallMIS in which {\em all} MISs are robust. Then, we turn our attention to the graphs that {\em admit} a robust MIS (\existsMIS). This class has a more complex structure; we give a partial characterization in terms of elementary graph properties, then a complete characterization by means of a (polynomial time) decision algorithm that accepts if and only if a robust MIS exists. This algorithm can be adapted to construct such a solution if one exists.

cs.DM↗

arXiv2018

Deterministic Leader Election Takes $Θ(D + \log n)$ Bit Rounds

Leader election is, together with consensus, one of the most central problems in distributed computing. This paper presents a distributed algorithm, called \STT, for electing deterministically a leader in an arbitrary network, assuming processors have unique identifiers of size $O(\log n)$, where $n$ is the number of processors. It elects a leader in $O(D +\log n)$ rounds, where $D$ is the diameter of the network, with messages of size $O(1)$. Thus it has a bit round complexity of $O(D +\log n)$. This substantially improves upon the best known algorithm whose bit round complexity is $O(D\log n)$. In fact, using the lower bound by Kutten et al. (2015) and a result of Dinitz and Solomon (2007), we show that the bit round complexity of \STT is optimal (up to a constant factor), which is a significant step forward in understanding the interplay between time and message optimality for the election problem. Our algorithm requires no knowledge on the graph such as $n$ or $D$, and the pipelining technique we introduce to break the $O(D\log n)$ barrier is general.

cs.DC↗

arXiv2018

Design Patterns in Beeping Algorithms: Examples, Emulation, and Analysis

We consider networks of processes which interact with beeps. In the basic model defined by Cornejo and Kuhn (2010), processes can choose in each round either to beep or to listen. Those who beep are unable to detect simultaneous beeps. Those who listen can only distinguish between silence and the presence of at least one beep. We refer to this model as $BL$ (beep or listen). Stronger models exist where the nodes can detect collision while they are beeping ($B_{cd}L$), listening ($BL_{cd}$), or both ($B_{cd}L_{cd}$). Beeping models are weak in essence and even simple tasks are difficult or unfeasible within. We present a set of generic building blocks (design patterns) which seem to occur frequently in the design of beeping algorithms. They include multi-slot phases: the fact of dividing the main loop into a number of specialised slots; exclusive beeps: having a single node beep at a time in a neighbourhood (within one or two hops); adaptive probability: increasing or decreasing the probability of beeping to produce more exclusive beeps; internal (resp. peripheral) collision detection: for detecting collision while beeping (resp. listening). Based on these patterns, we provide algorithms for a number of basic problems, including colouring, 2-hop colouring, degree computation, 2-hop MIS, and collision detection (in $BL$). The patterns make it possible to formulate these algorithms in a rather concise and elegant way. Their analyses are more technical; one of them improves significantly upon that of the best known MIS algorithm by Jeavons et al. (2016). Finally, inspired by a technique from Afek et al. (2013), our last contribution is to show that any Las Vegas algorithm relying on collision detection can be transposed into a Monte Carlo algorithm without collision detection at the cost of a logarithmic slowdown, which we prove is optimal.

cs.DC↗

arXiv2018

Finding Structure in Dynamic Networks

This document is the first part of the author's habilitation thesis (HDR), defended on June 4, 2018 at the University of Bordeaux. Given the nature of this document, the contributions that involve the author have been emphasized; however, these four chapters were specifically written for distribution to a larger audience. We hope they can serve as a broad introduction to the domain of highly dynamic networks, with a focus on temporal graph concepts and their interaction with distributed computing.

cs.DC↗

arXiv2017

Maintaining a Distributed Spanning Forest in Highly Dynamic Networks

Highly dynamic networks are characterized by frequent changes in the availability of communication links. These networks are often partitioned into several components, which split and merge unpredictably. We present a distributed algorithm that maintains a forest of (as few as possible) spanning trees in such a network, with no restriction on the rate of change. Our algorithm is inspired by high-level graph transformations, which we adapt here in a (synchronous) message passing model for dynamic networks. The resulting algorithm has the following properties: First, every decision is purely local---in each round, a node only considers its role and that of its neighbors in the tree, with no further information propagation (in particular, no wave mechanisms). Second, whatever the rate and scale of the changes, the algorithm guarantees that, by the end of every round, the network is covered by a forest of spanning trees in which 1) no cycle occur, 2) every node belongs to exactly one tree, and 3) every tree contains exactly one root (or token). We primarily focus on the correctness of this algorithm, which is established rigorously. While performance is not the main focus, we suggest new complexity metrics for such problems, and report on preliminary experimentation results validating our algorithm in a practical scenario.

cs.DC↗

arXiv2017

Efficiently Testing T-Interval Connectivity in Dynamic Graphs

Many types of dynamic networks are made up of durable entities whose links evolve over time. When considered from a {\em global} and {\em discrete} standpoint, these networks are often modelled as evolving graphs, i.e. a sequence of graphs ${\cal G}=(G_1,G_2,...,G_δ)$ such that $G_i=(V,E_i)$ represents the network topology at time step $i$. Such a sequence is said to be $T$-interval connected if for any $t\in [1, δ-T+1]$ all graphs in $\{G_t,G_{t+1},...,G_{t+T-1}\}$ share a common connected spanning subgraph. In this paper, we consider the problem of deciding whether a given sequence ${\cal G}$ is $T$-interval connected for a given $T$. We also consider the related problem of finding the largest $T$ for which a given ${\cal G}$ is $T$-interval connected. We assume that the changes between two consecutive graphs are arbitrary, and that two operations, {\em binary intersection} and {\em connectivity testing}, are available to solve the problems. We show that $Ω(δ)$ such operations are required to solve both problems, and we present optimal $O(δ)$ online algorithms for both problems. We extend our online algorithms to a dynamic setting in which connectivity is based on the recent evolution of the network.

cs.DS↗

arXiv2017

Robustness in Highly Dynamic Networks

We investigate a special case of hereditary property that we refer to as {\em robustness}. A property is {\em robust} in a given graph if it is inherited by all connected spanning subgraphs of this graph. We motivate this definition in different contexts, showing that it plays a central role in highly dynamic networks, although the problem is defined in terms of classical (static) graph theory. In this paper, we focus on the robustness of {\em maximal independent sets} (MIS). Following the above definition, a MIS is said to be {\em robust} (RMIS) if it remains a valid MIS in all connected spanning subgraphs of the original graph. We characterize the class of graphs in which {\em all} possible MISs are robust. We show that, in these particular graphs, the problem of finding a robust MIS is {\em local}; that is, we present an RMIS algorithm using only a sublogarithmic number of rounds (in the number of nodes $n$) in the ${\cal LOCAL}$ model. On the negative side, we show that, in general graphs, the problem is not local. Precisely, we prove a $Ω(n)$ lower bound on the number of rounds required for the nodes to decide consistently in some graphs. This result implies a separation between the RMIS problem and the MIS problem in general graphs. It also implies that any strategy in this case is asymptotically (in order) as bad as collecting all the network information at one node and solving the problem in a centralized manner. Motivated by this observation, we present a centralized algorithm that computes a robust MIS in a given graph, if one exists, and rejects otherwise. Significantly, this algorithm requires only a polynomial amount of local computation time, despite the fact that exponentially many MISs and exponentially many connected spanning subgraphs may exist.

cs.DC↗

arXiv2014

Shortest, Fastest, and Foremost Broadcast in Dynamic Networks

Highly dynamic networks rarely offer end-to-end connectivity at a given time. Yet, connectivity in these networks can be established over time and space, based on temporal analogues of multi-hop paths (also called {\em journeys}). Attempting to optimize the selection of the journeys in these networks naturally leads to the study of three cases: shortest (minimum hop), fastest (minimum duration), and foremost (earliest arrival) journeys. Efficient centralized algorithms exists to compute all cases, when the full knowledge of the network evolution is given. In this paper, we study the {\em distributed} counterparts of these problems, i.e. shortest, fastest, and foremost broadcast with termination detection (TDB), with minimal knowledge on the topology. We show that the feasibility of each of these problems requires distinct features on the evolution, through identifying three classes of dynamic graphs wherein the problems become gradually feasible: graphs in which the re-appearance of edges is {\em recurrent} (class R), {\em bounded-recurrent} (B), or {\em periodic} (P), together with specific knowledge that are respectively $n$ (the number of nodes), $Δ$ (a bound on the recurrence time), and $p$ (the period). In these classes it is not required that all pairs of nodes get in contact -- only that the overall {\em footprint} of the graph is connected over time. Our results, together with the strict inclusion between $P$, $B$, and $R$, implies a feasibility order among the three variants of the problem, i.e. TDB[foremost] requires weaker assumptions on the topology dynamics than TDB[shortest], which itself requires less than TDB[fastest]. Reversely, these differences in feasibility imply that the computational powers of $R_n$, $B_Δ$, and $P_p$ also form a strict hierarchy.

cs.DC↗

arXiv2014

Testing Temporal Connectivity in Sparse Dynamic Graphs

We address the problem of testing whether a given dynamic graph is temporally connected, {\it i.e} a temporal path (also called a {\em journey}) exists between all pairs of vertices. We consider a discrete version of the problem, where the topology is given as an evolving graph ${\cal G}=\{G_1,G_2,...,G_{k}\}$ whose set of vertices is invariant and the set of (directed) edges varies over time. Two cases are studied, depending on whether a single edge or an unlimited number of edges can be crossed in a same $G_i$ (strict journeys {\it vs} non-strict journeys). In the case of {\em strict} journeys, a number of existing algorithms designed for more general problems can be adapted. We adapt one of them to the above formulation of the problem and characterize its running time complexity. The parameters of interest are the length of the graph sequence $k=|{\cal G}|$, the maximum {\em instant} density $μ=max(|E_i|)$, and the {\em cumulated} density $m=|\cup E_i|$. Our algorithm has a time complexity of $O(kμn)$, where $n$ is the number of nodes. This complexity is compared to that of the other solutions: one is always more costly (keep in mind that is solves a more general problem), the other one is more or less costly depending on the interplay between instant density and cumulated density. The length $k$ of the sequence also plays a role. We characterize the key values of $k, μ$ and $m$ for which either algorithm should be used. In the case of {\em non-strict} journeys, for which no algorithm is known, we show that some pre-processing of the input graph allows us to re-use the same algorithm than before. By chance, these operations happens to cost again $O(kμn)$ time, which implies that the second problem is not more difficult than the first.

cs.DS↗

arXiv2014

Characterizing Topological Assumptions of Distributed Algorithms in Dynamic Networks

Besides the complexity in time or in number of messages, a common approach for analyzing distributed algorithms is to look at the assumptions they make on the underlying network. We investigate this question from the perspective of network dynamics. In particular, we ask how a given property on the evolution of the network can be rigorously proven as necessary or sufficient for a given algorithm. The main contribution of this paper is to propose the combination of two existing tools in this direction: local computations by means of graph relabelings, and evolving graphs. Such a combination makes it possible to express fine-grained properties on the network dynamics, then examine what impact those properties have on the execution at a precise, intertwined, level. We illustrate the use of this framework through the analysis of three simple algorithms, then discuss general implications of this work, which include (i) the possibility to compare distributed algorithms on the basis of their topological requirements, (ii) a formal hierarchy of dynamic networks based on these requirements, and (iii) the potential for mechanization induced by our framework, which we believe opens a door towards automated analysis and decision support in dynamic networks.

cs.DC↗

arXiv2014

Un algorithme de test pour la connexité temporelle des graphes dynamiques de faible densité

We address the problem of testing whether a dynamic graph is temporally connected, i.e. a temporal path ({\em journey}) exists between all pairs of vertices. We consider a discrete version of the problem, where the topology is given as an evolving graph $\G=\{G_1,G_2,...,G_{k}\}$ in which only the set of (directed) edges varies. Two cases are studied, depending on whether a single edge or an unlimited number of edges can be crossed in a same $G_i$ (strict journeys {\it vs} non-strict journeys). For strict journeys, two existing algorithms designed for other problems can be adapted. However, we show that a dedicated approach achieves a better time complexity than one of these two algorithms in all cases, and than the other one for those graphs whose density is low at any time (though arbitrary over time). The time complexity of our algorithm is $O(kμn)$, where $k=|\G|$ is the number of time steps and $μ=max(|E_i|)$ is the maximum {\em instant} density, to be contrasted with $m=|\cup E_i|$, the {\em cumulated} density. Indeed, it is not uncommon for a mobility scenario to satisfy, for instance, both $μ=o(n)$ and $m=Θ(n^2)$. We characterize the key values of $k, μ$ and $m$ for which our algorithm should be used. For non-strict journeys, for which no algorithm is known, we show that a similar strategy can be used to answer the question, still in $O(kμn)$ time.

cs.DS↗

arXiv2013

Distributed Maintenance of Anytime Available Spanning Trees in Dynamic Networks

We address the problem of building and maintaining distributed spanning trees in highly dynamic networks, in which topological events can occur at any time and any rate, and no stable periods can be assumed. In these harsh environments, we strive to preserve some properties such as cycle-freeness or the existence of a root in each tree, in order to make it possible to keep using the trees uninterruptedly (to a possible extent). Our algorithm operates at a coarse-grain level, using atomic pairwise interactions in a way akin to recent population protocol models. The algorithm relies on a perpetual alternation of \emph{topology-induced splittings} and \emph{computation-induced mergings} of a forest of spanning trees. Each tree in the forest hosts exactly one token (also called root) that performs a random walk {\em inside} the tree, switching parent-child relationships as it crosses edges. When two tokens are located on both sides of a same edge, their trees are merged upon this edge and one token disappears. Whenever an edge that belongs to a tree disappears, its child endpoint regenerates a new token instantly. The main features of this approach is that both \emph{merging} and \emph{splitting} are purely localized phenomenons. In this paper, we present and motivate the algorithm, and we prove its correctness in arbitrary dynamic networks. Then we discuss several implementation choices around this general principle. Preliminary results regarding its analysis are also discussed, in particular an analytical expression of the expected merging time for two given trees in a static context.

cs.DC↗

arXiv2012

Expressivity of Time-Varying Graphs and the Power of Waiting in Dynamic Networks

In infrastructure-less highly dynamic networks, computing and performing even basic tasks (such as routing and broadcasting) is a very challenging activity due to the fact that connectivity does not necessarily hold, and the network may actually be disconnected at every time instant. Clearly the task of designing protocols for these networks is less difficult if the environment allows waiting (i.e., it provides the nodes with store-carry-forward-like mechanisms such as local buffering) than if waiting is not feasible. No quantitative corroborations of this fact exist (e.g., no answer to the question: how much easier?). In this paper, we consider these qualitative questions about dynamic networks, modeled as time-varying (or evolving) graphs, where edges exist only at some times. We examine the difficulty of the environment in terms of the expressivity of the corresponding time-varying graph; that is in terms of the language generated by the feasible journeys in the graph. We prove that the set of languages $L_{nowait}$ when no waiting is allowed contains all computable languages. On the other end, using algebraic properties of quasi-orders, we prove that $L_{wait}$ is just the family of regular languages. In other words, we prove that, when waiting is no longer forbidden, the power of the accepting automaton (difficulty of the environment) drops drastically from being as powerful as a Turing machine, to becoming that of a Finite-State machine. This (perhaps surprisingly large) gap is a measure of the computational power of waiting. We also study bounded waiting; that is when waiting is allowed at a node only for at most $d$ time units. We prove the negative result that $L_{wait[d]} = L_{nowait}$; that is, the expressivity decreases only if the waiting is finite but unpredictable (i.e., under the control of the protocol designer and not of the environment).

cs.DC↗

arXiv2012

Building Fastest Broadcast Trees in Periodically-Varying Graphs

Delay-tolerant networks (DTNs) are characterized by a possible absence of end-to-end communication routes at any instant. Still, connectivity can generally be established over time and space. The optimality of a temporal path (journey) in this context can be defined in several terms, including topological (e.g. {\em shortest} in hops) and temporal (e.g. {\em fastest, foremost}). The combinatorial problem of computing shortest, foremost, and fastest journeys {\em given full knowledge} of the network schedule was addressed a decade ago (Bui-Xuan {\it et al.}, 2003). A recent line of research has focused on the distributed version of this problem, where foremost, shortest or fastest {\em broadcast} are performed without knowing the schedule beforehand. In this paper we show how to build {\em fastest} broadcast trees (i.e., trees that minimize the global duration of the broadcast, however late the departure is) in Time-Varying Graphs where intermittent edges are available periodically (it is known that the problem is infeasible in the general case even if various parameters of the graph are know). We address the general case where contacts between nodes can have arbitrary durations and thus fastest routes may consist of a mixture of {\em continuous} and {\em discontinuous} segments (a more complex scenario than when contacts are {\em punctual} and thus routes are only discontinuous). Using the abstraction of \tclocks to compute the temporal distances, we solve the fastest broadcast problem by first learning, at the emitter, what is its time of {\em minimum temporal eccentricity} (i.e. the fastest time to reach all the other nodes), and second by building a {\em foremost} broadcast tree relative to this particular emission date.

cs.DC↗

arXiv2012

Time-Varying Graphs and Dynamic Networks

The past few years have seen intensive research efforts carried out in some apparently unrelated areas of dynamic systems -- delay-tolerant networks, opportunistic-mobility networks, social networks -- obtaining closely related insights. Indeed, the concepts discovered in these investigations can be viewed as parts of the same conceptual universe; and the formal models proposed so far to express some specific concepts are components of a larger formal description of this universe. The main contribution of this paper is to integrate the vast collection of concepts, formalisms, and results found in the literature into a unified framework, which we call TVG (for time-varying graphs). Using this framework, it is possible to express directly in the same formalism not only the concepts common to all those different areas, but also those specific to each. Based on this definitional work, employing both existing results and original observations, we present a hierarchical classification of TVGs; each class corresponds to a significant property examined in the distributed computing literature. We then examine how TVGs can be used to study the evolution of network properties, and propose different techniques, depending on whether the indicators for these properties are a-temporal (as in the majority of existing studies) or temporal. Finally, we briefly discuss the introduction of randomness in TVGs.

cs.DC↗

arXiv2012

On the Feasibility of Maintenance Algorithms in Dynamic Graphs

Near ubiquitous mobile computing has led to intense interest in dynamic graph theory. This provides a new and challenging setting for algorithmics and complexity theory. For any graph-based problem, the rapid evolution of a (possibly disconnected) graph over time naturally leads to the important complexity question: is it better to calculate a new solution from scratch or to adapt the known solution on the prior graph to quickly provide a solution of guaranteed quality for the changed graph? In this paper, we demonstrate that the former is the best approach in some cases, but that there are cases where the latter is feasible. We prove that, under certain conditions, hard problems cannot even be approximated in any reasonable complexity bound --- i.e., even with a large amount of time, having a solution to a very similar graph does not help in computing a solution to the current graph. To achieve this, we formalize the idea as a maintenance algorithm. Using r-Regular Subgraph as the primary example we show that W[1]-hardness for the parameterized approximation problem implies the non-existence of a maintenance algorithm for the given approximation ratio. Conversely we show that Vertex Cover, which is fixed-parameter tractable, has a 2-approximate maintenance algorithm. The implications of NP-hardness and NPO-hardness are also explored.

cs.CC↗

arXiv2011

Time-Varying Graphs and Social Network Analysis: Temporal Indicators and Metrics

Most instruments - formalisms, concepts, and metrics - for social networks analysis fail to capture their dynamics. Typical systems exhibit different scales of dynamics, ranging from the fine-grain dynamics of interactions (which recently led researchers to consider temporal versions of distance, connectivity, and related indicators), to the evolution of network properties over longer periods of time. This paper proposes a general approach to study that evolution for both atemporal and temporal indicators, based respectively on sequences of static graphs and sequences of time-varying graphs that cover successive time-windows. All the concepts and indicators, some of which are new, are expressed using a time-varying graph formalism.

cs.SI↗

arXiv2010

Construction et maintien d'une forêt couvrante dans un réseau dynamique

In this work we introduce the principles of an algorithm that constructs and maintains a spanning forest in a mobile telecommunication network-a MANET. The algorithm is based on the random walk of a token and is entirely decentralized. A probability analysis is performed when the network is static. Then we show that performances can be slightly enhanced when adding a memory process in the walk on the token.

cs.NI↗