Matteo Maria Andreozzi

Ns2Voip++, an enhanced module for VoIP simulations

In the last ten years, many circuit-switched networks for voice have been replaced with packet-switched ones. Hence, simulating Voice over IP has become of paramount importance in assessing the performance of a network. However, a sound performance analysis should be carried out in conditions which are as close as possible to a real deployment. In this paper we present enhancements to ns2voip [1], a module for simulating realistic VoIP traffic with the ns2 simulator. In detail, we add new features, i.e. a correlated model for packet generation in a two-way conversation and implementing a set of realistic playout buffers to simulate the behavior of the receiver. Our code is available at http://cng1.iet.unipi.it/wiki/index.php/Ns2voip.

Flexible scheduling for Real-Time services in High-Speed Packet Access cellular networks

This paper presents the HYbrid Channel-Aware and Real-Time scheduler (HY-CART) for UMTS High-Speed Packet Access networks. The latter works for both downlink and uplink scheduling, with slightly different embodiments due to the different technology constraints. Its basic principle is to select users based on their channel conditions when the deadlines are far ahead: as the deadlines approach, they become progressively more important. Furthermore, resources for scheduled users are allocated sparingly, i.e. enough to serve packets whose deadline is actually near, so as to make room for a possibly larger number of simultaneous users, and keeping into account the user channel state. Simulative evaluation shows that HY-CART performs better than the previous work, allowing more users to be served with the same resources.

Throughput-optimal resource allocation in LTE-Advanced with distributed antennas

Distributed antennas are envisaged for LTE-Advanced deployments in order to improve the coverage and increase the cell throughput. The latter in turn depends on how resources are allocated to the User Equipments (UEs) at the MAC layer. In this paper we discuss how to allocate resources to UEs so as to maximize the cell throughput, given that UEs may receive from several antennas simultaneously. We first show that the problem is both NP-hard and APX-hard, i.e. no polynomial-time algorithm exists that approximates the optimum within a constant factor. Hence, we propose and evaluate two polynomial-time heuristics whose complexity is feasible for practical purposes. Our simulative analysis shows that, in practical scenarios, the two heuristics are highly accurate.

LTE and LTE-Advanced evaluation through innovative simulation tool

The study of next generation wireless networks as LTE or LTE Advanced is not trivial. We developed a complete system level simulator for studying such emerging technology and we also propose a solution for one of the core points of LTE-A: the scheduling of Relay nodes wireless links activation.

OptiMOS: Optimal MOS-based scheduling of downlink voice flows in point-to-multipoint access networks

We introduce OptiMOS, a method for selecting deadlines for the scheduling of downlink voice flows in a centralized-scheduling point-to-multipoint network. These include cellular networks such as LTE, UMTS HSDPA, or WiMAX, and WLANs with centralized control. OptiMOS selects the deadline of each packet so as to maximize the expected Mean Opinion Score (MOS) for the talkspurt it belongs to. It requires that absolute time is both included in RTP timestamps at the source and checked for scheduling decisions, in a cross-layer approach. OptiMOS can be employed in conjunction with any deadline-based scheduling algorithm, and does not require a specific playout buffering scheme at the receiver. Simulations of an LTE system show a performance improvement in terms of MOS for the VoIP flows, good tolerance to relatively large synchronization errors and qualitatively similar results with different playout buffers.

Effective scheduling of real-time traffic in HSUPA

Scheduling of real-time uplink flows in HSUPA is hampered by the fact that the NodeB does not know the size and dead-lines of the uplink packets, as the User Equipment (UE) only reports coarse-grained Scheduling Information (SI) advertising the overall backlog. In this paper, we describe how the NodeB can use the sequence of SI to reconstruct a UE queue, using affordable computations. This makes it possible to use scheduling algorithms employing packet sizes and deadlines in HSUPA, with obvious performance benefits. We present a general Virtual Queueing framework, that can be employed with any kind of traffic, and a specialized version of it for periodic (e.g., voice) traffic.