Gérard Hébuterne

Simultaneity in discrete-time single server queues with Bernoulli inputs

Abstract In discrete-time systems, arrivals and departures occur simultaneously. The present work focuses on the scheduling of arrivals and departures in discrete-time queues with Bernoulli arrivals and general independent service times. Both finite and infinite capacity queues are analysed. The waiting-room management policy takes care of the simultaneity problem. Two waiting-room management policies, AF (Arrivals First) and DF (Departures First) are singled out and their influence on the derivation of the state probabilities is investigated. Furthermore, the outside observers distribution is shown to be identical to the arriving customers distribution in the AF queue and a formula relating both distributions is provided for the DF queues.

A space priority queuing mechanism for multiplexing ATM channels

Abstract In an ATM network, it can be necessary to protect some cells against rejection when entering finite-capacity buffers. “Space priority” is an access mechanism intended to provide such a protection by allowing them to take the place of less important cells in case of congestion. We study an M/G/1 queue using this operating mode and we compute loss probabilities for both flows. The model is then compared with the usual finite capacity queue in order to qualify space priority: reduction in buffer size, and increase in the admissible load for a given Grade of Service.

Relation between states observed by arriving and departing customers in bulk systems

An important and well known theorem of Queueing Theory establishes the equality of state distributions at arrival and departure epochs, provided that the system state varies by unit jumps. We generalize this result to a class of bulk systems (arrival occur in groups of random size, or departures in groups of fixed size), for which we give a relation between state distributions at arrival and departure instants, under quite general conditions.

On the capacity of multi-cell UMTS

In this paper, we focus on the capacity of a multicell UMTS system. We first determine an upper bound on the other cell interference and obtain novel expressions for the SIR and powers for both the uplink and the downlink. The former is an asynchronous CDMA system, often using one of three types of receivers : matched filter, minimum mean-square error (MMSE) and decorrelator. In the latter, the SIR depends on the distance between the user and the base station. We show that the inter-cell interference alters the capacity of the system for all kinds of receivers, though the effective bandwidth expressions remain identical to the ones obtained in the single-cell case. As an application, we show how to use this analytical model to develop a new efficient connection admission control (CAC) algorithm.

Traffic engineering in a multipoint-to-point network

The need to guarantee quality-of-service (QoS) to multimedia applications leads to a tight integration between the routing and forwarding functions in the Internet. multiprotocol label switching tries to provide a global solution for this integration. In this context, multipoint-to-point (m2p) networks appear as a key architecture since they provide a cheaper way to connect edge nodes than point-to-point connections. M2p networks have been mainly studied for their load balancing ability. In this paper, we go a step further: we propose and evaluate a traffic management scheme that provides deterministic QoS guarantees for multimedia sources in an m2p network. We first derive an accurate upper bound on the end-to-end delay in an m2p architecture based on the concept of additivity. Broadly speaking, an m2p network is additive if the maximum end-to-end delay is equal to the sum of local maximum delays. We then introduce two admission control algorithms for additive networks: a centralized algorithm and a distributed algorithm and discuss their complexity and their scalability.

Shaping self-similar traffic at access of optical network

The focus of this paper is shaping of self-similar traffic at the access of an optical node. Due to the lack of optical memories, we suggest a combined exploitation of electronic memories in the edge of the optical network by means of shaping and propose two novel shaping algorithms that dynamically shape the incoming traffic so as to meet Quality of Service (QoS) requirements while keeping with the optical core network constraints. The first algorithm follows a deterministic approach and is based on the service curves theory. It dynamically, on an interval basis, adjusts the shaping parameters to the token bucket described self-similar traffic. The second, alternative shaping algorithm follows a stochastic approach. It blindly shapes the traffic using a jumping observation window.

Deterministic End-to-End Delay Bounds in an Accumulation Network

In this paper, we focus on the determination of end-to-end delay bounds for FIFO accumulation networks with leaky bucket constrained sources, including IP and ATM network cases. We determine an upper bound for the overall end-to-end delay and prove its accuracy to approximate the exact maximum end-to-end delay for accumulation networks of any size. This is achieved through an original trajectory analysis approach. Numerical studies further illustrate this point. This extends our previous results for the two-server and proves that this bound may be used as an accurate criterion for a CAC algorithm providing a deterministic QoS guaranteed service.

Measurement-based admission control in UMTS multiple cell case

In this paper, we develop an efficient call admission control (CAC) algorithm for UMTS systems. We first introduce the expressions that we developed for signal-to-interference (SIR) for both uplink and downlink, to obtain a novel CAC algorithm that takes into account, in addition to SIR constraints, the effects of mobility, coverage as well as the wired capacity in the LMTS terrestrial radio access network (UTRAN). for the uplink, and the maximal transmission power of the base station, for the downlink. As of its implementation, we investigate the measurement-based approach as a means to predict future, both handoff and new, call arrivals and thus manage different priority levels. Compared to classical CAC algorithms, our CAC mechanism achieves better performance in terms of outage probability and QoS management.

Dynamic Shaping for Self-Similar Traffic Using Network Calculus

The focus of this paper is the shaping of self-similar traffic at the access of an optical node. We propose a novel algorithm that dynamically shapes the incoming traffic, based on service curves equations, in order to meet the optical nodes constraints in terms of buffer size or delay. We first estimate arrival parameters within various time intervals in order to make the incoming traffic fit into a token bucket traffic specification (Tspec) format. We then derive the shaping parameters based on deterministic service curves. Those shaping parameters vary dynamically according to the Tspec of every time window. We eventually set those parameters back into the original model in order to meet some QoS constraints at the optical network level.

Mapping of Performance Metrics Between IP and ATM

End-to-end QoS requires mapping of QoS between heterogeneous subnetworks of the end-to-end path. For this mapping to be accurate, we argue, in this paper, that it should be based on the performance metrics. We propose a method to map loss and delay between IP, particularly integrated services (intserv), and ATM, using deterministic service curves. First, we consider a single IP flow and show our original method for mapping of backlog and delay between IP and ATM. Second, we extend our mapping to the case of the three intserv classes served by a single ATM switch implementing a priority-based discipline. Third, we focus on the cases of non-conformant traffic and congested systems. For our mapping method to cover those cases, we introduce loss in the non-lossy service curves approach, through two different ways. Extensive numerical simulations yield qualitative as well as quantitative mapping of loss and delay between intserv and ATM.