Jean-Marie Garcia

Aggregate Traffic Models for VoIP Applications

Performance evaluation of VoIP applications requires reliable traffic models. Although good traffic models for unitary VoIP applications do exist, the simulation of thousands of connections is very slow making simulation studies very heavy. Aggregate traffic models are needed to achieve faster simulations without losing accuracy. In this paper, we characterize the superposition of homogenous and heterogeneous VoIP applications under heavy and light traffic intensities. We suggest simple aggregate traffic models for VoIP applications and show the performance limits of these models in network simulation environment

Analytic approximation of the jitter incurred by CBR traffics in IP networks

In this paper, an analytic approximation is derived for the end-to-end delay-jitter incurred by a periodic traffic with constant packet size. The single node case is considered first. It is assumed that the periodic traffic is multiplexed with a background packet stream under the FCFS service discipline. The processes governing the packet arrivals and the packet sizes of the background traffic are assumed to be general renewal processes. A very simple analytical approximation is derived and its accuracy is assessed by means of event-driven simulations. This approximation is then extended to the multiple node case yielding a simple analytical approximation of the end-to-end jitter. This approximation is shown to be fairly accurate in the light to moderate traffic conditions typically encountered in IP core networks.

Bandwidth Allocation and Session Scheduling using SIP

Session Initiation Protocol (SIP) is a new signaling protocol designed to establish multimedia sessions in telecommunication networks. In this paper, we suggest the extension of SIP functionalities to coordinate QoS mechanisms deployed in IP networks, and especially in DiffServ domain. Indeed, the interaction between small and big TCP sessions may have dramatic consequences on small TCP sessions. Hence, we use SIP to achieve QoS management on a session basis, in which the over all activity of the user during the session is considered. The suggested mechanisms deal with two issues: first, session scheduling based on session duration and/or volume, and second bandwidth allocation on a per-flow basis using equivalent bandwidth estimation techniques. The proposed mechanisms are implemented in the SIP proxy server as QoS management algorithms, and they are validated by simulations.

A Model for Telephone Networks and Its Use for Routing Optimization Purposes

The problem studied here concerns the modeling of call blocking in telephone networks. From the usual assumptions such as exponential arrivals and holding time, lost call cleared, the state of the network is described by a finite Markov chain. From the transition probabilities of this process are derived the differential equations associated with the average occupancy of all trunk groups. These traffic equations are simplified by considering independence of blocking for trunk groups in series. The blocking probabilities are estimated using fictitious offered traffic and the Erlang B formula. Such representation takes into account peaky or smooth traffic characteristics. We develop this one-moment model for routing policies such as load sharing and overflow routing. Performances of the model are given in comparison to the solution of the exact Markov chain model or the results of Monte Carlo simulation. Finally, an application to routing optimization and network dimensioning is treated.

A Simple Formula for End-to-End Jitter Estimation in Packet-Switching Networks

In this paper, an analytic approximation is derived for the end-to-end delay-jitter incurred by a periodic traffic with constant packet size. It is assumed that the periodic traffic is multiplexed with a background packet stream under the FCFS service discipline in each queue along the path to its destination. The processes governing the packet arrivals and the packet sizes of the background traffics are assumed to be general renewal processes. A very simple analytical approximation is derived and its accuracy is assessed by means of event-driven simulations.

Integration of streaming and elastic traffic: a fixed point approach

We present a fixed point approach to evaluate the quality of service of streaming traffic multiplexed with elastic traffic in multi-service networks. First, we handle elastic traffic and streaming traffic separately, and then we derive a general fixed point formulation integrating both types of traffic in best effort networks. Then, we extend the application of this formulation to multi-service networks where priorities and bandwidth sharing schemes can be applied to different flows. Our approach is mainly oriented towards very large scale networks where traditional simulation techniques are not scalable, and where a large number of flows have to be evaluated in reasonable time. We assess the accuracy of our approach by means of event-driven simulations.

Fanout Inference from Link Count

The traffic matrix is the fundamental input data in network planning, simulation and traffic engineering. However, it is often unknown and its direct measurement with devices such as Netflow is a too heavy process for large high-speed networks. The estimation of the traffic matrix from link counts appears as the best alternative approach. Recent works assume that prior information do not allow alone an accurate estimation of the traffic matrix so they are using Netflow measures to calibrate their model. These models assume spatial and temporal relations between different instants of measure. We show in this paper that we can obtain similar error rates without this costly calibration phase thanks to a spatial and temporal relation introduced in K. Papagiannaki et al. (2004)

Internet Traffic Modeling

Traffic modeling is an important tool for performance evaluation of networks. Simple models like Poisson process have been discredited in modeling Internet traffic. At the same time, characterization studies of Internet traffic revealed the presence of different types of correlations in Internet traffic. In this paper, we explore the use of M|G|infin process in capturing traffic correlations. We use real traffic traces to construct M|G|infin traffic models for both video and Internet background traffic. We evaluate the traffic generated both statistically and in real network simulations

Hybrid Approach for Energy Aware Management of Multi-cloud Architecture Integrating user Machines

The arrival and development of remotely accessible services via the cloud has transfigured computer technology. However, its impact on personal computing remains limited to cloud-based applications. Meanwhile, acceptance and usage of telephony and smartphones have exploded. Their sparse administration needs and general user friendliness allows all people, regardless of technology literacy, to access, install and use a large variety of applications. We propose in this paper a model and a platform to offer personal computing a simple and transparent usage similar to modern telephony. In this model, user machines are integrated within the classical cloud model, consequently expanding available resources and management targets. In particular, we defined and implemented a modular architecture including resource managers at different levels that take into account energy and QoS concerns. We also propose simulation tools to design and size the underlying infrastructure to cope with the explosion of usage. Functionalities of the resulting platform are validated and demonstrated through various utilization scenarios. The internal scheduler managing resource usage is experimentally evaluated and compared with classical methodologies, showing a significant reduction of energy consumption with almost no QoS degradation.

Modeling Internet Traffic: Performance Limits

Traffic modeling is an important tool for performance evaluation of networks. Over the last decade traffic characterization studies revealed the presence of different types of correlations in Internet traffic. In this paper we present traffic modeling methodology by aggregates of bytes using the M/G/infin process for capturing traffic correlations. We use recent traffic traces to construct M/G/infin traffic models, and we evaluate the generated traffic both statistically and by simulation in network environment. We show limitations of bytes aggregates models, and we conclude by propositions to obtain more accurate models