• Nidhi Hegde (Nokia, France)
  Tools for the analysis of content dissemination in social networks
  Information propagation in networks has been studied for decades. In the past, such "rumour spread" has been based on a model where a source emits an information which then spreads in the network according to push, pull, or hybrid mechanisms. The goal in such work has been to characterize the rate of spread across the whole network. More recently, information propagation from the perspective of social networks has been studied, where a more realistic model of multiple sources of information, and multiple 'types' of information is considered. In such a model, each node in the network has resource constraints and must choose on how to relay the information.
  In this overview, we will go over some recent work on information spread in social networks, covering analysis and algorithms for efficient dissemination. In particular, we will consider distributed algorithms for optimal information propagation, and we review game-theoretic tools used in the characterisation of such distributed control problems in this model and its variants.

• Patrick Loiseau (EURECOM Institute, Sophia Antipolis, France)
  Strategic resource allocation in adversarial environments
  Allocation of resources is a well-known problem that is often solved by optimization techniques. In adversarial environments, however, the objective function (or payoff) depends on on how an adversary allocates his resources. Examples of such situations are numerous and include allocation of security defenses to different targets (where the payoff depends on how an attacker allocates his attack resources) and allocation of advertisement/lobbying resources to customers/voters (where the payoff depends on how a competitor allocates his resources). In such scenarios, the resource allocation problem becomes a game.
  In this talk, we review strategic resource allocation games. The fundamental model of strategic resource allocation is the Colonel Blotto game, proposed by Borel in 1921. Two players allocate an exogenously given amount of resources to a fixed number of battlefields with given values. Each battlefield is then won by the player who allocated more resources to it, and each player maximizes the aggregate value of battlefields he wins. Despite its apparent simplicity, the Colonel Blotto game is still unresolved in its most general form. We review existing solutions and briefly mention some interesting variants of this game.

Tutoriel: Asymptotic analysis techniques for performance evaluation

• Rudesindo Núñez Queija (Univ. of Amsterdam et CWI, Amsterdam)
  For many queueing systems exact analysis of performance measures such as queue lengths, waiting times and sojourn time is often out of reach. Also, average values may not even be the most informative measures to describe a system's performance, but one may rather be interested in performance quantiles for example. For such cases a wide range of asymptotic techniques are available that may serve to develop suitable approximations and provide valuable insights. In this course we will briefly outline several such techniques (large deviations and tail asymptotics, fluid and diffusion limits, perturbation analysis, heavy traffic limits) and illustrate them on queueing models such as GPS queues, DPS queues, and bandwidth-sharing networks.
  The main part of the tutorial will focus on a more detailed discussion of two particular techniques:
  - Heavy-tailed asymptotics: We will explain the fundamental difference with light tailed asymptotics ("conspiracy" versus "disaster" scenarios) and illustrate several analysis techniques that one may resort to in obtaining asymptotically accurate estimates, including analytic asymptotics, probabilistic bounds and coupling arguments.
  - Perturbation analysis (in particular time-scale separation): analyzing Markovian queueing networks as multi-dimensional Markov processes may be notoriously difficult. One abstraction is to isolate the behavior of a single queue, and capture the influence of other queues in what is called the random environment. As the random environment changes state, the queue can move from one mode of operation to another (for example from lightly loaded conditions to overloaded conditions and back). Perturbation techniques provide approximations when the state changes of the random environment occur on a much faster or much smaller time scale than the queueing dynamics.
  
  

Tutoriel-Démo: Outils logiciels du projet ANR MARMOTE

• Jean-Michel Fourneau (Université de Versailles Saint Quentin)
  Titre: Xborne: génération, comparaison, résolution de chaînes de Markov
• Jean-Marc Vincent (Université Joseph Fourier)
  Titre: Psi3: modélisation par événements et simulation exacte de chaînes de Markov
• Alain Jean-Marie (INRIA Sophia-Antipolis Méditerranée)
  Titre: marmoteCore: créer et résoudre des chaînes de Markov en C++
  
  

Démo: NEST - A Demonstration on Network Modeling and Simulation

• A. Al Sheikh (QoS Design, Toulouse)
  In this technical demonstration we will highlight key functionalities of NEST, QoS Design's software suites aimed at network operators. We will first demonstrate NEST IP/MPLS as we focus on defining an IP/MPLS network and associated parameters and protocols. Features such as tracing traffic flow routes throughout the network interfaces, editing access populations and technologies, and interpreting main simulation and performance evaluation results will be presented. Afterwards, we will overview other NEST software concerning optical networks and network supervision.