Jean-Loup Guillaume

Bipartite structure of all complex networks

The analysis and modelling of various complex networks has received much attention in the last few years. Some such networks display a natural bipartite structure: two kinds of nodes coexist with links only between nodes of different kinds. This bipartite structure has not been deeply studied until now, mainly because it appeared to be specific to only a few complex networks. However, we show here that all complex networks can be viewed as bipartite structures sharing some important statistics, like degree distributions. The basic properties of complex networks can be viewed as consequences of this underlying bipartite structure. This leads us to propose the first simple and intuitive model for complex networks which captures the main properties met in practice.

Temporal reachability graphs

While a natural fit for modeling and understanding mobile networks, time-varying graphs remain poorly understood. Indeed, many of the usual concepts of static graphs have no obvious counterpart in time-varying ones. In this paper, we introduce the notion of temporal reachability graphs. A (tau,delta)-reachability graph is a time-varying directed graph derived from an existing connectivity graph. An edge exists from one node to another in the reachability graph at time t if there exists a journey (i.e., a spatiotemporal path) in the connectivity graph from the first node to the second, leaving after t, with a positive edge traversal time tau, and arriving within a maximum delay delta. We make three contributions. First, we develop the theoretical framework around temporal reachability graphs. Second, we harness our theoretical findings to propose an algorithm for their efficient computation. Finally, we demonstrate the analytic power of the temporal reachability graph concept by applying it to synthetic and real-life datasets. On top of defining clear upper bounds on communication capabilities, reachability graphs highlight asymmetric communication opportunities and offloading potential.

Clustering in p2p exchanges and consequences on performances

We propose here an analysis of a rich dataset which gives an exhaustive and dynamic view of the exchanges processed in a running eDonkey system. We focus on correlation in term of data exchanged by peers having provided or queried at least one data in common. We introduce a method to capture these correlations (namely the data clustering), and study it in detail. We then use it to propose a very simple and efficient way to group data into clusters and show the impact of this underlying structure on search in typical P2P systems. Finally, we use these results to evaluate the relevance and limitations of a model proposed in a previous publication. We indicate some realistic values for the parameters of this model, and discuss some possible improvements.

Relevance of massively distributed explorations of the internet topology: qualitative results

Internet maps are generally constructed using the traceroute tool from a few sources to many destinations. It appeared recently that this exploration process gives a partial and biased view of the real topology, which leads to the idea of increasing the number of sources to improve the quality of the maps. In this paper, we present a set of experiments we have conducted to evaluate the relevance of this approach. It appears that the statistical properties of the underlying network have a strong influence on the quality of the obtained maps, which can be improved using massively distributed explorations. Conversely, some statistical properties are very robust, and so the known values for the Internet may be considered as reliable. We validate our analysis using real-world data and experiments, and we discuss its implications.

Comparison of failures and attacks on random and scale-free networks

It appeared recently that some statistical properties of complex networks like the Internet, the World Wide Web or Peer-to-Peer systems have an important influence on their resilience to failures and attacks. In particular, scale-free networks (i.e. networks with power-law degree distribution) seem much more robust than random networks in case of failures, while they are more sensitive to attacks. n nIn this paper we deepen the study of the differences in the behavior of these two kinds of networks when facing failures or attacks. We moderate the general affirmation that scale-free networks are much more sensitive than random networks to attacks by showing that the number of links to remove in both cases is similar, and by showing that a slightly modified scenario for failures gives results similar to the ones for attacks. We also propose and analyze an efficient attack strategy against links.

Statistical analysis of a p2p query graph based on degrees and their time-evolution

Despite their crucial impact on the performances of p2p systems, very few is known on peers behaviors in such networks. We propose here a study of these behaviors in a running environment using a semi-centralised p2p system (eDonkey). To achieve this, we use a trace of the queries made to a large server managing up to fifty thousands peers simultaneously, and a few thousands queries per second. We analyse these data using complex network methods, and focus in particular on the degrees, their correlations, and their time-evolution. Results show a large variety of observed phenomena, including the variety of peers behaviors and heterogeneity of data queries, which should be taken into account when designing p2p systems.

Efficient and Simple Encodings for the Web Graph

In this paper, we propose a set of simple and efficient methods based on standard, free and widely available tools, to store and manipulate large sets of URLs and large parts of the Web graph. Our aim is both to store efficiently the URLs list and the graph in order to manage all the computations in a computer central memory. We also want to make the conversion between URLs and their identifiers as fast as possible, and to obtain all the successors of an URL in the Web graph efficiently. The methods we propose make it possible to obtain a good compromise between these two challenges, and make it possible to manipulate large parts of the Web graph.

Multi-ego-centered communities in practice

We propose here a framework to unfold the ego-centered community structure of a given node in a network. The framework is not based on the optimization of a quality function, but on the study of the irregularity of the decrease of a proximity measure. It is a practical use of the notion of multi-ego-centered community and we validate the pertinence of the approach on benchmarks and a real-world network of wikipedia pages.

Studying Graph Dynamics Through Intrinsic Time Based Diffusion Analysis

Complex networks may be studied in various ways, e.g., by analyzing the evolutions of their topologies over time, and in particular of their community structures. In this paper, we focus on another type of dynamics, related to diffusion processes on these networks. Indeed, our work aims at characterizing network dynamics from the diffusion point of view, and reciprocally, it evaluates the impact of graph dynamics on diffusion. We propose in this paper an innovative approach based on the notion of intrinsic time, where the time unit corresponds to the appearance of a new link in the graph. This original notion of time allows us to somehow isolate the diffusion phenomenon from the evolution of the network. The objective is to compare the diffusion features observed with this intrinsic time concept from those obtained with traditional (extrinsic) time, based on seconds. The comparison of these time concepts is easily understandable yet completely new in the study of diffusion phenomena. We experiment our approach on three real datasets and show the promising results of intrinsic time-based diffusion analysis.