Daniel Kofman

A Stochastic Geometry Analysis of Dense IEEE 802.11 Networks

This paper presents a stochastic geometry model for the performance analysis and the planning of dense IEEE 802.11 networks. This model allows one to propose heuristic formulas for various properties of such networks like the probability for users to be covered, the probability for access points to be granted access to the channel or the average long term throughput provided to end-users. The main merit of this model is to take the effect of interferences and that of CSMA into account within this dense network context. This analytic model, which is based on Matern point processes, is partly validated against simulation. It is then used to assess various properties of such networks. We show for instance how the long term throughput obtained by end-users behaves when the access point density increases. We also briefly show how to use this model for the planning of managed networks and for the economic modeling of unplanned networks.

DPS queues with stationary ergodic service times and the performance of TCP in overload

In a previous paper, BonaId and Roberts (June 2001) studied non-persistent TCP connections in transient overload conditions, under the assumption that all connections have the same round-trip times. In this paper our goal is to develop theoretical tools that will enable us to relax this assumption and obtain explicit expressions for the rate of growth of the number of connections at the system, the rate at which TCP connections leave the system, as well as the time needed for the completion of a connection. To that end, we model the system as a DPS (discriminatory processor sharing) system which we analyze under very mild assumptions on the probability distributions related to different classes of arrivals: we only assume that the arrival rates of connections exist, and that the amount of information transmitted during a connection of a given type forms a stationary ergodic sequence. We then proceed to obtain explicit expressions for the growth rate of the number of connections at the DPS system for several specific probability distributions. We check through simulations the applicability of our queueing results for modeling TCP connections sharing a bottleneck.

A Smart Parking Lot Management System for Scheduling the Recharging of Electric Vehicles

In this paper, we propose a centralized electric vehicles (EVs) recharge scheduling system for parking lots using a realistic vehicular mobility/parking pattern focusing on individual parking lots. We consider two different types of EV based on their mobility/parking patterns: 1) regular EVs; and 2) irregular EVs. An extensive trace-based vehicular mobility model collected from the Canton of Zurich is used for the regular EVs, and a probabilistic pattern built on top of this trace is used for modeling the behavior of irregular EVs. To the extent of our knowledge, this is the first EV charging scheduling study in the literature that takes into account a realistic vehicular mobility pattern focusing on individual parking lots. We compare the performance of our proposed system with two well-known basic scheduling mechanisms, first come first serve and earliest deadline first, with regard to two objective functions: 1) maximizing the total parking lot revenue; and 2) maximizing the total number of EVs fulfilling their requirements. Comparison results show that our proposed system outperforms well-known basic scheduling mechanisms with regards to both objectives. Parking lots managing the recharging of a high number of EVs will greatly benefit from using such recharge scheduling systems in the context of smart cities.

Self organizing hierarchical multicast trees and their optimization

Multicast routing protocols suitable for wide-area networks are being developed for the Internet. Protocols based on hierarchical trees appear to be well suited for their superior scalability and flexibility. We show how to construct a class of hierarchical multicast trees and we analyze their performances. This study gives insight into how the chosen hierarchical structure impacts the tree performance with respect to network resource consumption. Optimal structures which minimize resource consumption are deduced, which allows for simple dimensioning rules, such as how many hierarchical levels should be used. A stochastic geometric approach turned out to be well adapted for this study. This approach leads to explicit expressions for the average tree cost, as a function of the hierarchical clustering, from which the optimal tree structure can then be easily deduced.

Discrete time queues with delayed information

We study the behavior of a single-server discrete-time queue with batch arrivals, where the information on the queue length and possibly on service completions is delayed. Such a model describes situations arising in high speed telecommunication systems, where information arrives in messages, each comprising a variable number of fixed-length packets, and it takes one unit of time (a slot) to transmit a packet. Since it is not desirable to attempt service when the system may be empty, we study a model where we assume that service is attempted only if, given the information available to the server, it is certain that there are messages in the queue. We characterize the probability distribution of the number of messages in the queue under some general stationarity assumptions on the arrival process, when information on the queue size is delayedK slots, and derive explicit expressions of the PGF of the queue length for the case of i.i.d. batch arrivals and general independent service times. We further derive the PGF of the queue size when information onboth the queue length and service completion is delayedK=1 units of time. Finally, we extend the results to priority queues and show that when all messages are of unit length, thecμ rule remains optimal even in the case of delayed information.

A Scalable IoT Service Search Based on Clustering and Aggregation

With the Internet of Things (IoT), we are facing a proliferation of connected devices distributed over a collection of geographical locations and offering new services to users referred to as IoT services. These services are highly heterogeneous and their number is growing fast. To facilitate IoT service search, some solutions rely on the use of Semantic Web technologies to create homogenous service descriptions to reason on and support accurate and flexible search. However, most existing approaches are centralized and fail to scale with respect to the number of advertised services. To address this scalability issue, we present in this paper a distributed, semantic-based IoT service search system. It relies on a hierarchical network of semantic gateways that represent geographical locations and host semantic service descriptions. To answer a service query, our approach is based on request matching and forwarding though the use of routing tables and similarity thresholds. Based on defined metrics, recursive clustering and information aggregation are performed over the hierarchy of gateways to construct the routing tables. We prove theoretically that our method ensures an accurate search of all matching services within the system. Experimental results show that our method enables to reduce greatly search cost with comparison to a centralized approach and has an acceptable search path length.

Optimal Virtual Topology Design using Bus-Label Switched Paths

Despite an observable trend in reducing the number of layers in core and metropolitan area networks, still the optimal design of multi-layer networks (like IP over WDM) remains an important issue as a means to reduce CAPEX and OPEX. In multi-layer networks, usually a connection oriented transport network (like WDM) is used to deploy a layout for optimal transport of traffic. In todays networks, the layouts are based on point to point LSPs. We will show here that by using a novel concept named bus-LSPs, we can achieve some significant savings for the operator, both in terms of CAPEX and OPEX. We also introduce a heuristic algorithm to design optimal layouts, and provide a quantitative evaluation

Packet based load sharing schemes in MPLS networks

By adding a connection oriented layer to legacy IP routing and forwarding, the MPLS architecture provides new mechanism to perform traffic engineering in IP networks. Current MPLS request for comments and drafts have identified MPLS protection switching and load sharing as attractive applications of these mechanisms, overcoming the limitations of legacy IP routing. Existing load balancing IP routing protocols and architectures (ECMP, OSPF-OMP, etc.) have limited load sharing capabilities to avoid loops (e.g. only for equal cost paths for ECMP), usually based on fixed administrative metrics. Given the connection nature of MPLS networks, arbitrary load sharing algorithms can be performed without leading to loops. In this paper we propose and evaluate a family of load sharing schemes that take into account the nature of the aggregated traffic as well as the link capacities and administrative lengths. We propose cost functions that reflect/penalize LSP congestion, using overflow estimates and we derive optimality conditions. For example, for equal length paths (in the sense of legacy administrative metrics) under particular conditions, the optimal load share is based on the sole bottleneck capacity of the LSP: the optimal load share splits traffic according to the pragmatic rule of proportional capacities regardless of the aggregated traffic, as long as the stability constraints are respected. However, the statistical properties of the traffic (thus the cost functions) give important feedback and stochastic bounds when not in the optimal case, either in the event of a failure or when the real bottleneck capacity differs from the nominal one. We illustrate our schemes with simple but significant examples.

Polling systems with station breakdowns

Abstract None of the works in the literature on Polling Systems deals with the important phenomenon of failing nodes. This paper addresses this issue. We assume that there are two modes of station breakdowns: node-dependent and general. Each node is subject to a random failure process and a failure of a node during a visit of the server causes the latter to move to the next station. We carry performance analyses of both the Gated and the Exhaustive service regimes and derive probability generating functions, Laplace transforms and means of variables such as queue sizes at polling and at arbitrary instants, durations of busy periods, number of jobs left behind at servers station-departure moments, and cycle times. The analysis of the Exhaustive case is achieved via an explicit solution of an infinite set of linear equations, where the unknowns are the state-dependent joint transforms of two key characteristics: the number of arrivals to a station during a busy period starting with a given number of jobs, and the number of jobs successfully served during that period. Stability conditions are indicated for each regime.

Efficient semantic-based IoT service discovery mechanism for dynamic environments

The adoption of Service Oriented Architecture (SOA) and semantic Web technologies in the Internet of Things (IoT) enables to enhance the interoperability of devices by abstracting their capabilities as services and to enrich their descriptions with machine-interpretable semantics. This facilitates the discovery and composition of IoT services. The increasing number of IoT services, their dynamicity and geographical distribution require mechanisms to enable scalable and efficient discovery. We propose in this paper a semantic based IoT service discovery system that supports and adapts to the dynamicity of IoT services. The discovery is distributed over a hierarchy of semantic gateways. Within a semantic gateway, we implement mechanisms to dynamically organize its content over time, in order to minimize the discovery cost. Results show that our approach enables to maintain a scalable and efficient discovery and limits the number of updates sent to a neighboring gateway.