Giovanni Ciccarese

Current methods for calculating sustainability metrics of network links are based on fixed-rate estimations. We present a context-aware model of energy consumption, also providing an open-source tool that enables developers to use the Internet more efficiently. In a proof of concept, we collected and analyzed a two-month dataset covering 43 AWS and Microsoft endpoints.

Vehicular communications could be exploited for energy management of vehicles. We propose a system which provides that a vehicle estimates its future speed profile gathering status messages broadcasted by the surrounding vehicles and/or the infrastructure and inputting them in a traffic simulator used as a predictor. The system has been validated by simulation considering an urban scenario inspired to the Ecotekne campus at the University of Salento and a Manhattan scenario, very challenging in relation to the prediction of the speed profile. Simulation results have shown that the prediction error is quite low for the first scenario. In the Manhattan scenario, the error is quite high in case each vehicle limits itself to send messages only to its neighbours and does not transmit the information regarding its route. However, the error can be significantly reduced if route information is broadcasted and the infrastructure relays the messages transmitted by vehicles. The proposed system has been tested in the Ecotekne campus.

A number of Intelligent Flooding Schemes have been recently proposed in order to optimize message dissemination in Vehicular Ad Hoc Networks. In this paper we focus on a scheme which exploits a distributed timer-based contention mechanism for allowing only contention winners to forward the message and to suppress other potential forwarders. In particular, at each hop along the message propagation direction, potential forwarders wait for a time which is inversely proportional to their distance from the sender before rebroadcasting the packet; a potential forwarder is suppressed if it intercepts the packet rebroadcasted by another node during the waiting time. Performance of the scheme, in terms of message delivery ratio, delay and channel utilization, depends on the maximum waiting time (MaxWT) and on the maximum distance R allowed between the sender and a potential forwarder. This research work aims at evaluating an optimal setting of these parameters. We show that, although the values of MaxWT and R which maximize performance vary with traffic load and vehicle density, a dynamic algorithm is not essential. Then, we report how to set the values of the parameters to achieve a performance that is acceptable in the scenarios which we have considered.

Broadcasting will be extensively used in Vehicular Ad Hoc Networks (VANET), mainly for spreading out safety messages. A number of Intelligent Flooding Schemes (IFSs) have been recently proposed to optimize message dissemination. Some of them exploit a distributed timer-based contention mechanism in order to allow the most suitable nodes to forward the packet and to suppress other potential forwarders. This paper deals with those schemes which are based on a handshake similar to RTS/CTS. This helps in avoiding potential forwarders, which should be suppressed, to uselessly rebroadcast a data packet. We aimed at identifying main problems which these schemes may encounter in optimizing message dissemination. To this objective, an analysis of the events which can cause the failure of the forwarding election procedure has been carried out. The analysis has also been the starting point for devising some different design choices.

Recently, there has been much effort to integrate heterogeneous wireless technologies in order to make overlay networks able to provide mobile users with ubiquitous Internet access. Handover management is still one of the most challenging issues to be solved for seamless integration of wireless networks. In this paper, an algorithm for vertical handover with QoS guarantees between IEEE 802.16e and IEEE 802.11e networks, without computational complexity added on mobile nodes, has been proposed. It uses a decision mechanism based on the following metrics: the available bandwidth, the signal-to-noise ratio (SNR) and the access delay. Its characteristics are perfectly compatible with the guidelines specified in the IEEE 802.21 standard. The effectiveness of the defined algorithm has been evaluated, using the ns-2 tool, in terms of packet loss rate (PLR), satisfaction of minimum applications requirements and number of useless handovers. The performance of the defined algorithm has been compared with that obtained by using the algorithm suggested by IEEE 802.21 standard that exploits only the SNR in the decision procedure.

This paper proposes a novel algorithm which enables an ARQ protocol defined for IEEE 802.16e to vary the packet size depending on wireless channel behaviour. The algorithm adopts an analytical approach for dynamically evaluating the optimal packet size. The latter is derived from an expression of the ARQ protocol efficiency, obtained by exploiting a finite-state Markov error model which also considers Adaptive Modulation/Coding. The effectiveness of the designed algorithm in improving TCP performance has been evaluated.

In this paper, a packet size control algorithm, fully compliant with the IEEE 802.16e standard, has been proposed. This algorithm aims to optimize the performance of a data link ARQ protocol, trying to evaluate dynamically the packet size that maximizes the mean advancement of the sender window. It offers the possibility to maximize system throughput in every mobile scenario, even in presence of wireless time-variant channels. The effectiveness of the defined algorithm has been evaluated, through computer simulation, in terms of normalized TCP goodput. Simulation results have shown that the dynamic choice of the packet size plays an important role in maximizing TCP performance.

Mobile computing and security in Internet have become two of the most important challenges in the information technology world. Among wireless technologies which support user mobility, satellites will play a very important role. Unfortunately, TCP does not perform well on satellite channels because of the related high delay-bandwidth product, high bit error rate (BER) and burst errors mainly due to shadowing. In order to obtain improvements in TCP performance, an approach local to the satellite link could be adopted. In fact, recently, some studies showed the effectiveness of the use of a reliable data link protocol that is well suited to mobile geo-stationary satellite channel characteristics. However, the resultant more variable end-to-end delay causes, as a side effect, a new problem: the competition between retransmission mechanisms at transport layer and data link layer. The introduction of a suitable performance enhancing proxy (PEP) on the satellite gateway has proved to be a viable technique to face the problem. However, the compatibility with the usage of network layer security mechanisms, such as IPSec, should be pursued. This paper just deals with the definition of an IPSec-aware TCP PEP, called secure PEP (SPEP), that is completely transparent to IPSec architecture. The performances of SPEP have been validated in the context of Web application by using network simulator tool. The simulation results, in terms of Web-page download mean time, have shown that the proposed solution provides substantial improvements in TCP performance.