Section: New Results

Distributed algorithms for dynamic networks

Participants : Luciana Bezerra Arantes [correspondent] , Rudyar Cortes, Raluca Diaconu, Jonathan Lejeune, Olivier Marin, Sébastien Monnet, Franck Petit [correspondent] , Karine Pires, Pierre Sens, Véronique Simon, Julien Sopena.

Nowadays, distributed systems are more and more heterogeneous and versatile. Computing units can join, leave or move inside a global infrastructure. These features require the implementation of dynamic systems, that is to say they can cope autonomously with changes in their structure in terms of physical facilities and software. It therefore becomes necessary to define, develop, and validate distributed algorithms able to managed such dynamic and large scale systems, for instance mobile ad hoc networks, (mobile) sensor networks, P2P systems, Cloud environments, robot networks, to quote only a few.

Efficiency in such environments requires specialised protocols, providing features such as fault or heterogeneity tolerance, scalability, quality of service, and self-*. Our approach covers the whole spectrum from theory to experimentation. We design algorithms, prove them correct, implement them, and evaluate them in simulation, using OMNeT++ or PeerSim, and on large-scale real platforms such as Grid'5000. The theory ensures that our solutions are correct and whenever possible optimal; experimental evidence is necessary to show that they are relevant and practical.

Within this thread, we have considered a number of specific applications, including massively multi-player on-line games (MMOGs) and peer certification.

We have obtained results both on fundamental aspects of distributed algorithms and on specific emerging large-scale applications.

We study various key topics of distributed algorithms: mutual exclusion, failure detection, data dissemination and data finding in large scale systems, self-stabilization and self-* services.

Self-Stabilization.

We have also approached fault tolerance through self-stabilization. Self-stabilization is a versatile technique to design distributed algorithms that withstand transient faults.

In [43] , we proposed a silent self-stabilizing leader election algorithm (SSLE, for short) for bidirectional connected identified networks of arbitrary topology. Starting from any arbitrary configuration, SSLE converges to a terminal configuration, where all processes know the ID of the leader, this latter being the process of minimum ID. Moreover, as in most of the solutions from the literature, a distributed spanning tree rooted at the leader is defined in the terminal configuration. This algorithm is written in the locally shared memory model. It assumes the distributed unfair daemon, the most general scheduling hypothesis of the model. Our algorithm requires no global knowledge on the network (such as an upper bound on the diameter or the number of processes, for example). We showed that its stabilization time is in $\Theta \left(n^3\right)$ steps in the worst case, where $n$ is the number of processes. Its memory requirement is asymptotically optimal, i.e., $\Theta (logn)$ bits per processes. Its round complexity is of the same order of magnitude — i.e., $\Theta \left(n\right)$ rounds — as the best existing algorithm designed with similar settings. To the best of our knowledge, this was the first self-stabilizing leader election algorithm for arbitrary identified networks that is proven to achieve a stabilization time polynomial in steps. By contrast, we show that the previous best existing algorithm designed with similar settings stabilizes in a non polynomial number of steps in the worst case.

We have also implemented SSLE in a high-level simulator to empirically evaluate its average performances. Experimental results tend to show that its worst case in terms of rounds ($\Theta (3n+D)$ rounds) is rare.

Dynamic Distributed Systems

The first key challenge in understanding highly dynamic networks consists in developing appropriate models that are as close as possible to the phenomena that one wishes to capture. This requires the use of a formalism sufficiently expressive to formulate complex temporal properties. Recently, a vast collection of concepts, formalisms, and models has been unified in a framework called Time-Varying Graphs (TVG) (A. Casteigts, P. Flocchini, W. Quattrociocchi, and N. Santoro, Time-varying graphs and dynamic networks, International Journal of Parallel, Emergent and Distributed Systems 27(5):387-408, 2012), which are represented as time-ordered sequences of graphs defined over a fixed set of nodes. A hierarchy of classes over TVG has been described, mainly depending on properties related to connectivity and recurrence of dynamic. Such an hierarchy is an interesting tool for study computability issues. As an example, if one is able to prove an impossibility result in a class of the hierarchy with strong properties, then this impossibility result also holds in any class of the hierarchy with (strictly) weaker properties. In this context, we provide a generic framework to prove impossibility results in this model [45] . This framework helps to formally prove classical arguments about convergence of sequence of time-varying graphs used to build counter-examples. We apply this generic framework to the study of covering problems (such as minimal dominating set and maximal matching) in the context of time-varying graphs. We obtain a characterization of the weakest topology assumption that makes these problems computable. We also propose a general time complexity measure since time-varying graph model lacks so far of such a definition.

Swarm of Mobile Robots

Swarm of autonomous mobile sensor devices (or, robots) recently emerged as an attractive issue in the study of dynamic distributed systems permits to assess the intrinsic difficulties of many fundamentals tasks, such as exploring or gathering in a discrete space. We consider autonomous robots that are endowed with visibility sensors (but that are otherwise unable to communicate) and motion actuators. Those robots must collaborate to solve a collective task, namely exclusive perpetual exploration, despite being limited with respect to input from the environment, asymmetry, memory, etc. The area to be explored is modeled as a graph and the exclusive perpetual exploration task requires every possible vertex to be visited infinitely often by every robot, with the additional constraint that no two robots may be present at the same node at the same time or may concurrently traverse the same edge of the graph.

In [28] , we presented and implemented a generic method for obtaining all possible protocols for a swarm of mobile robots operating in a particular discrete space, namely an anonymous rings. Our method permits to discover new protocols that solve the problem, and to assess specific optimization criteria (such as individual coverage, visits frequency, etc.) that are met by those protocols. To our best knowledge, this was the first attempt to mechanize the discovery and fine-grained property testing of distributed mobile robot protocols.