This paper presents DroneOPERA, an ns-3 simulation module to model the energy consumption of on-board computing in energy-constrained unmanned aerial vehicles (UAVs) operating in Internet of Drones (IoD) networks. As drone networks are increasingly integrated with IoT and cloud services, energy-efficient processing becomes a key challenge, especially in applications requiring real-time analytics under strict power budgets. DroneOPERA extends ns-3 with a configurable model that couples UAV mobility, battery dynamics, and CPU workloads associated with on-board tasks. By processing data locally instead of offloading raw information to external servers, DroneOPERA enables the study of privacy-preserving mission designs, where communication is limited to compact, processed results. The module exposes APIs to define mission profiles, processing workloads, and hardware configurations, and computes the resulting energy expenditure at fine time granularity. We demonstrate the capabilities of DroneOPERA through a representative use case, in an agriculture scenario, highlighting the trade-offs between task complexity, flight time, and privacy-preserving on-board processing.

Modern Software-Defined Wide Area Network (SDWAN) deployments are required to manage traffic over heterogeneous underlay networks while meeting stringent Quality of Service (QoS) requirements. In scenarios where multiple branches share overlay resources, independent tunnel selection decisions often lead to congestion and degraded performance. Existing approaches lack coordination mechanisms to handle the dynamic interactions between agents competing for shared resources. This paper presents a Multi-Agent Reinforcement Learning (MARL) framework for distributed overlay selection in SD-WANs. Each branch is modeled as an autonomous agent that learns routing policies through interaction with the network environment. To account for the mutual impact of decisions across branches, we adopt the Centralized Training with Decentralized Execution (CTDE) paradigm, enabling agents to learn globally consistent behaviors while preserving scalability at inference. To encourage cooperative policies, we introduce a $\lambda$-weighted reward shaping mechanism that balances local QoS goals with global resource fairness. We evaluate our approach using both PPO and DQN algorithms in a simulated SD-WAN environment. The findings highlight the necessity of MARL in addressing resource contention and ensuring equitable shared overlay utilization.

In rural areas such as national parks, forests, and mountains, recognizing and classifying wildlife is crucial for monitoring animals that may pose risks to crops and human safety. Deep learning provides the most accurate approach due to its dynamic adaptability. However, it is also highly energy intensive. Given the constraints of rural environments, including the use of extreme edge devices and limited power availability, this study employs distributed computing to maximize battery life while enabling the execution of DL tasks. The core concept is Split Computing, specifically applied to the YOLOv8m and YOLOv10 models, which have been provided by researchers as a highly effective solution for wildlife classification. This approach dynamically reallocates different model components across available boards, specifically the NVIDIA Jetson Orin Nano, to optimize energy consumption. This research investigates various model distribution configurations across the head, backbone, and neck components, assessing whether video stream compression between the camera and the computing boards affects energy consumption and network load. Our findings suggest that the choice of model splitting configuration significantly impacts energy efficiency, bandwidth consumption, and computational load distribution.

This study investigates the application of deep convolutional neural networks (CNNs) for lung disease classification from chest X-ray images, with a focus on deployment in resourceconstrained settings on low-power edge devices. Six state-of-the-art CNN architectures, including ResNet101, DenseNet201, MobileNetV3-Large, EfficientNetV2-B0, InceptionResNetV2, and Xception, were evaluated on a base dataset and an augmented dataset to assess their generalization capacity. Data augmentation showed varied impacts: ResNet101 exhibited significant improvement in validation accuracy with augmented data, indicating enhanced generalization, while models like InceptionResNetV2 showed limited improvement, suggesting underfitting. Additionally, inference performance on the Raspberry Pi 4 Model B demonstrated the impact of model complexity on speed and latency. MobileNetV3-Large and EfficientNetV2-B0 emerged as the most suitable models for real-time deployment, balancing accuracy with computational efficiency, whereas deeper models incurred higher latency, making them less practical for edge deployment. These findings underscore the importance of data augmentation, model selection, and inference efficiency in optimizing CNN-based lung disease classifiers for real-world healthcare applications. Future work will extend the evaluation to include heterogeneous edge devices, further exploring scalability and energy efficiency for real-time deployment in rural clinical settings with limited resources.

The integration of machine learning on microcontrollers at the extreme edge of the network is gaining significance for various Internet of Things applications. This paradigm shift enables real-time data processing and decisionmaking directly on the devices, thereby reducing latency and reliance on cloud connectivity. Nevertheless, due to the computational and memory limitations of microcontrollers, training and deploying a neural network using existing libraries such as TinyML remains impractical for some devices. This paper proposes an approach where the model is trained in the cloud, with only the forward phase directly implemented on the microcontroller. We leverage the LittleBits platform for its user-friendly interface and suitability for educational purposes, aiming to introduce machine learning concepts to non-expert users. Our case study focuses on temperature prediction in smart greenhouses, demonstrating the practical utility of machine learning in timely interventions for plant disease prevention and optimal growth conditions. This work not only illustrates the feasibility of deploying machine learning on resource-constrained devices but also emphasizes the potential of LittleBits in making novel technologies accessible and comprehensible to a broader audience.

The digital divide remains one of the biggest challenges of the modern era, limiting access to information and global connectivity for populations in remote or rural regions. The doctoral research goes in the direction of an integrated approach to address the divide in the view of 6G. In particular, the integration of Non-Terrestrial Networks (NTN) and Terrestrial Networks (TN) is envisioned as a key part of the future 6G systems. Exploiting the Software-Defined Wide Area Network (SD-WAN) technology, we aim to propose an architecture for an adaptive and resilient network that can integrate satellite technologies such as Low Earth Orbit (LEO) satellite networks. Moreover, we present some results of using Reinforcement Learning (RL) in the tunnel selection problem for SD-WAN. In order to assess the performance of the LEO satellite, we are also collecting and analyzing real measurements using a Starlink connection. Finally, we propose some open questions in the field of integrating NTN/TN with SD-WAN.

Smart agriculture is experiencing rapid transformation through innovative Agritech solutions. Robotic Arms (RAs) and Sensor-equipped Wearable Gloves (SWGs) are critical tools for performing high-precision tasks in controlled environments like hydroponic greenhouses. Digital Twin (DT) technology further enhances these systems, enabling real-time monitoring, dynamic adaptation, and predictive maintenance, thus ensuring optimal efficiency and cost savings. This paper investigates the integration of DT technology into a hydroponic greenhouse to facilitate remote control of a RA using a SWG. We have developed a DT framework that synchronizes physical and digital environments for real-time motion planning and monitoring. Our system architecture aims to demonstrate the potential of DTs in agriculture, allowing tasks to be performed in three modes: Real-to-Digital (updating the DT with real-world data), Digital-to-Real (executing commands from the digital model in the physical world), and Digital-to-Digital (simulating scenarios to optimize performance). Experimental validation highlights the system's performance, focusing on communication and computing delays and their impact on real-time operations.

Climate change poses significant challenges, particularly in agriculture, where extreme weather events demand more efficient and resilient systems, such as smart greenhouses. These controlled environments require predictive solutions to optimize conditions such as temperature and humidity, which are critical for crop growth. While neural networks are widely employed for climate prediction, their complexity presents a barrier to implementation on edge devices, which are characterized by limited computational resources. In this work, we propose Fuzzy-Augmented Neural Network (FANN), a novel approach based on fuzzy sets, applied in cascade to regressive neural network models, to reduce complexity and energy consumption in greenhouse microclimate classification and prediction. The methodology was tested on four edge devices, including microcontrollers and microprocessors. We compare our FANN approach with standard models (FFNN, BNN, SimpleRNN, GRU, LSTM), highlighting significant reductions in inference time, energy consumption, and memory usage. FANN also offers practical advantages, such as the ability to adapt classification by modifying fuzzification parameters without retraining the model, and the potential to parallelize computations for simultaneously classifying the microclimate of multiple crops. These features make the system flexible and optimal for practical applications in dynamic agricultural contexts.

Networked Music Performance (NMP) is one of the most challenging real-time applications in which musicians can play together using the internet without being physically together. Nowadays, the connection technology can have performance that guarantees an adequate Quality of Experience (QoE) for NMP. On the other hand, there are millions of musicians who live in remote and rural areas without access to high-speed network connections. The promising LEO satellite internet technology could fill this gap and democratize NMP in remote places. This paper aims to provide an analysis of Starlink’s capability to meet the stringent requirements of NMP. We present an analysis of network metrics that are relevant to NMP applications and assess the NMP software JackTrip in order to test a real NMP application with LEO satellite internet. The evaluation of the RTT, One-way delay, Packet loss, and jitter in different scenarios poses some challenges for the NMP scenario. We present the results of two live NMP jamming sessions with musicians, proposing an adapted QoE model that incorporates both objective and subjective metrics. With a tailored buffer strategy and the ability of a musician, we were able to obtain an adequate QoE, overcoming the challenges introduced by LEO satellite internet.

The agricultural sector is undergoing a significant transformation driven by technological advancements led by Agritech solutions. In this context, our study introduces the architecture of smart Hydroponic greenhouses. This work focuses on developing and testing a 3D sensing glove designed to empower users in remote control. Additionally, we developed a 3D model of the human hand to test the movement accuracy of the sensing glove. Latency measurements on serial, Bluetooth 4.0 interfaces and a proposed Network Architecture are conducted to assess real-time control performance. Our architectural proposal aims to control agricultural machinery seamlessly within hydroponic greenhouses, enhancing precision and user experience. The study lays the foundation for future advancements in remote agricultural operations and smart greenhouse management.

Surgical instrument detection and tracking is essential in surgical procedures. The presence and proper use of surgical instruments is crucial for patient safety and overall patient outcomes. Most existing deep learning-based surgical instrument detection systems require high-end GPUs or cloud services for deployment. However, such systems are not feasible in resource-constrained or remote areas where access to expensive and bulky equipment is limited. To overcome these challenges, we propose a cost-effective and sustainable solution for deep learning based surgical instrument detection. Our approach is to detect surgical instruments from images using pretrained models on a Raspberry Pi. We compare two lightweight, pretrained deep learning models, EfficientNet and MobileNet, on their accuracy and efficiency for this task. These models are optimized for edge devices with limited computational resources. We evaluate the tradeoff between model accuracy and computational efficiency on a benchmark dataset. We discuss the strengths and limitations of each model, as well as their implications for developing surgical instrument detection applications for edge devices. This work can enable the selection of effective deep learning models for real-time inference on edge devices, and facilitate the development of efficient and cost-effective healthcare solutions.

The continuous growth of advanced services integrating processing, storage, real-time data exchange, and transmission capacities highlights the importance of Quality of Service (QoS) and Quality of Experience (QoE). In particular, it is very relevant for enterprises that rely on Wide Area Networks (WANs) for reliable and efficient communication between their headquarters and branches. With the evolution of new applications with stringent requirements, ensuring high-performing WANs is a critical priority for businesses looking to remain competitive and provide a seamless customer experience. This paper explores using Rein-forcement Learning (RL) algorithms on Software-Defined Wide Area Network (SD-WAN) to improve QoS and reduce costs. SD-WAN allows for the dynamic reconfiguration of network devices in real-time, better meeting network measurements and service requirements. By leveraging self-learning techniques such as RL, which exploits feedback mechanisms, we can improve network availability by automatically routing traffic over existing network technologies. We also compare two SD-WAN topology scenarios, including direct WAN connections between Customer Premises Equipment (CPEs) within enterprise premises and CPEs used as peering points for traffic routing. Our approach shows promising results regarding network performance and cost-effectiveness, which can benefit businesses looking to improve their network infrastructure.

Recent impressive growth of AI applications in the most diversified heterogeneous domains is largely motivated by the availability of hardware accelerators used from the backstage of data centers (such as TPU, Tensor Processing Units, or VPUs, Visual Processing Units) to the far edge of embedded devices equipped with DPUs and Deep Learning Processing Units. High level toolchains for a more friendly usability of these platform had similar relevance in the process. In this paper we considered edge devices that provide an essential contribution for the deployment of “distributed intelligence” and are used typically at the gateway, CPE or Edge computing level. One of the typical assumptions is that Field Programmable Gate Array (FPGA) are far more expensive - with respect to power consumption - than legacy SBCs (single board computers). The main contribution of the paper is a fair comparison (at the same clock frequency and with the same main CPU) of processing time and power consumption of two different boards used for deep neural network classification. We will highlight the relevance of classification speed with respect to common KPIs adopted to compare the performances of automatic classification such as Loss, Precision, Recall, etc. This will be particularly relevant in the challenging domains of hardware accelerated real time control loops to provide distributed intelligence at the application level but also at the inner functions of emerging networking architectures.

Quality of Service (QoS) and Quality of Experience (QoE) are the most relevant requirements for new advanced services comprising integrated cooperation between processing, storage, sensing, and transmission capabilities. As reliance on technology grows, businesses increasingly turn to Wide Area Networks (WANs) connectivity solutions to ensure reliable and efficient communication between their locations and cloud data centers. With new applications with stringent requirements, QoS and QoE have become critical priorities for companies looking to remain competitive and provide a seamless customer experience. As a result, many organizations require reliable and high-performing WANs to effectively transmit critical data between their branches and cloud data centers. We examine Multi-Protocol Label Switching (MPLS technology), which is the historical choice for WANs but has significant disadvantages in cost and performance. Then, we analyze the novel Software-Defined Wide Area Network (SD-WAN) technology that allows for real-time flexible, dynamic reconfiguration of network devices to meet network measurements and service requirements. Furthermore, we introduce a first work of my Ph. D. that examines the use of a Reinforcement Learning algorithm in a new SD-WAN topology scenario in which we consider both direct WAN connections between Customer Premises Equipment (CPEs) located within enterprise premises and CPEs used as peering points for traffic routing. This paper presents our initial analysis of these technologies and the main ideas guiding me in my Ph. D. program.

The deep convolutional neural networks are widely used in medical image classification tasks. In some cases, they have outperformed physicians and achieved significant results. Unlike natural images, medical image dataset are very hard to collect, because they are protected by the privacy regulations to preserve patient's anonymity and requires a great deal of professional expertise to label them. However, because of the easier access and availability of high-performance computational resources, leveraging deep neural networks to detect diseases is becoming increasingly popular and common practice among healthcare researchers. As a result, considerable amount of energy is consumed to find an optimal and effective solution, which has a huge impact on our environment and contributes to global warming to some level.To address these challenges and reduce the carbon footprint caused by the deep learning practitioners, we attempted to combine the advantages of both federated learning and transfer learning for the medical image classification task in our study. Our findings suggest that federated transfer learning could be an useful technique to minimize computational costs and energy efficient, while maintaining privacy and addressing the problem of data scarcity. Moreover, this approach can be applied to solve other healthcare related tasks.

Modern multi-core servers are powerful enough to process multi-gigabit live packet streams on the network data plane. However, in most cases network programmers must build their applications from scratch, by implementing both the interfaces towards the lower hardware level and the proper mechanisms for parallel programming. Data Stream Processing (DaSP) frameworks have recently emerged as promising approaches to overcome the above issues and to let programmers simply focus on the logic of the application to develop. However, DaSP platforms are generally not designed for the networking domain, in terms of both performance and functions. In this paper, we selected the WindFlow DaSP framework and built suitable extensions to attach multiple (accelerated) packet sources of data to it. We then implemented a simple monitoring application on top of WindFlow and carried out stress tests with synthetic and real traffic. The results prove that performance scale linearly with the processing cores so that the application was able to process the whole amount of live data up to nearly 20 Gbps rate.

We present a vertical integration of a real-time Industrial Internet of Things environment with Cloud computing functionalities. We designed our testbed to implement self-adaptive wings for motorbikes using pressure values sensed in specific locations of the motorbike as input data collected exploiting open-hardware devices called OpenMotes which communicate through a low-power, delay-constrained wireless network based on the 6LoWPAN protocol stack. Our real-time on-board unit elaborates the data, and it computes the new angle of attack of two wings. The total response time of the system is in the order of 100 ms and meets the real-time requirements that constrains our scenario. Finally, we integrate our system with Cloud functionalities that we use for storing the acquired data on a time-series database.

The Internet of Things (IoT) significantly extends the attack surface of the Internet, making the use of an Anomaly-based Intrusion Detection System of paramount importance. Despite in the last years big research efforts have focused on the application of Deep Learning techniques to attack detection, an ultimate real-time solution, able to provide a high detection rate with an acceptable false alarm rate while processing raw network traffic in real time, has still to be identified. For this reason, in this paper we propose an Intrusion Detection System that, leveraging on probabilistic data structures and Deep Learning, is able to process in real time the traffic collected in a backbone network, offering almost optimal detection performance and low false alarm rate. Indeed, the extensive experimental tests, run to validate and evaluate our system, confirm that, with a proper parameter setting, we can achieve about 90% of detection rate, with an accuracy of 0.871.

Time-slotted channel hopping (TSCH) mode in the IEEE 802.15.4-2015 standard provides ultrahigh reliability and ultralow power consumption to sensor devices. The key feature of TSCH is the scheduling of time slots and frequencies, which falls outside the current standards. In this paper, we focus on throughput maximizing and max-min fair scheduling problems in a centralized TSCH networks. At first, a polynomial time algorithm for the throughput maximizing scheduling problem is proposed. We proceed to investigate and deliberate on some instances of the problem with their combinatorial properties. Second, a novel auction-based scheduling algorithm that uses a first-price sealed-bid auction mechanism is presented for the throughput maximizing problem. Simulation results show that the proposed algorithm obtains a close throughput performance to the optimal one obtained through CPLEX with a much lower complexity. Moreover, we propose a novel heuristic for the max-min fair scheduling problem and demonstrate its performance through extensive simulations in terms of the total throughput and fairness varying the number of nodes, frequencies, and antennas. Simulation results indicate the effectiveness of the proposed algorithm and its close performance to the optimal solution.

Due to the impressive demand for communication-intensive applications deployed in the cloud, cross-layer orchestration solutions for Cloud data centers (DCs) are essential for the effective usage of both network and cloud resources while addressing service requirements and user expectations. In this regard, the software-defined networking (SDN) can play a key role thanks to programmable network control operations and to a more effective inter-working between network controllers and cloud management platforms. On the other hand, due to the relevant size of Cloud DCs, a significant evaluation of SDN-based orchestration solutions requires testing environments with proper scales, and significant level of fidelity. This paper discusses the opportunities of using virtual testbeds to carry out Cloud-related experimentations while addressing with significant scale requirements of DC infrastructures. First, we present a set of results of a test campaign we carried out using the Fed4FIRE experimental facility aiming at a comprehensive performance evaluation of an SDN-based orchestration solution for Cloud DCs. Second, we provide a comparison between virtual testbeds and other testing environments, i.e., laboratory testbeds, simulators and emulators, in terms of offered advantages, and effectiveness in experimentations. Moreover, we discuss the appropriateness and advisability of using virtualized testbeds for research on novel SDN and Cloud orchestration solutions at scale, by highlighting advantages of virtual testbeds in comparison with laboratory, simulation and emulation testing facilities in terms of effectiveness of experimentations and accuracy of results.

In this paper, we introduce an autonomous system for aerodynamic flow control for motorcycle based on the Internet of Things (IoT) paradigm. The architecture we propose adapts dynamically the flows at the traveling conditions, in order to obtain an improvement of performance and vehicle stability. In our architecture, we deploy a group of sensors on the top surface of the wings to sense the air pressure. We design a centralized on-board unit that computes a new wing angle of attack according to the data received from the sensors. The on-board unit includes a local information database which represents its knowledge: it stores both the data gathered by the sensors and the fluid dynamics model used to compute and adjust the angle of attack. The on-board database is periodically updated transmitting all the measurements gathered from the sensors to a High-Performance Cloud Data Center (DC) which executes a parallel version of Computational Fluid Dynamics (CDF) algorithms, computes the updated model, and transmits the processed information back to the on-board unit. We perform preliminary tests in the wind tunnel, and we show how the cooperation between IoT devices and DC can reduce the on-board unit computational effort.

The large availability of multi-gigabit network cards for commodity PCs requires network applications to potentially cope with high volumes of traffic. However, computation intensive operations may not catch up with high traffic rates and need to be run in parallel over multiple processing cores. As of today, the vast majority of network applications-e.g., monitoring and IDS systems-are still based on the pcap library interface which, unfortunately, does not provide the native multi-core support, even though the current underlying capture technologies do. This paper introduces a novel version of the pcap library for the Linux operating system that enables transparent application level parallelism. The new library supports fan-out operations for both multi-threaded and multi-process applications, by means of extended API as well as by a declarative grammar for configuration files, suitable for legacy applications. In addition, the library can transparently run on top of the standard Linux socket as well as on other accelerated active engines. Performance evaluation has been carried out on a multi-core architecture in pure capture tests and in more realistic use cases involving monitoring applications such as Tstat and Bro, with standard Linux socket as well as PF_RING and PFQ accelerated engines.

The widespread diffusion of mobile phones and the convergence of more and more services to the mobile devices allow the service providers to collect a huge corpus of information about the behaviour of the users (e.g., location and activity), “transforming” the cellular network into a highly distributed and pervasive sensor network. In this paper, we aim at demonstrating how, by using simple statistical tools, these data can be used to obtain information, often apparently unrelated to the mobile phone usage. To this aim, we have analysed the CDRs collected by Telecom Italia in December 2013 in Milan (Italy) and made available by the “Open Big Data” project, to infer information about the behaviour of foreign people in Milan. Such analysis clearly highlights the potential of applying big data analysis techniques to the mobile providers data.

This paper introduces a feasibility study of a low cost and power effective wireless acoustic sensor network for shooter localization. Currently deployed sensors for this application use a time domain gunshot signal analysis and high sampling rates, in the order of MS/s. We investigate an alternative Short Time Fourier Transform with a lower sampling rate (250 kS/s). We focus on a single channel single sensor approach for shooter ranging and we provide new experimental data. We implement a centralized gateway on top of the open and programmable board named ZedBoard. Our approach localizes the shooter with a distance error that ranges between roughly half and three meters, so the results are encouraging for further studies.

The evolution of commodity PCs towards multi-core processing platforms equipped with high-speed network interfaces makes them reasonable and cost effective targets for the implementation of generic network functions. In addition, the availability of software accelerated I/O frameworks provides a convenient ground for running a broad variety of applications, from simple software switches to more complex network systems, with near hardware-class performance and the flexibility of a software approach. Most network functions can be implemented by composing a set of elementary operations into processing pipelines to be run on top of multiple processing cores. In this framework, maintaining the flow consistency is crucial to enable stateful operations in the processing pipelines. This paper presents Enif-Lang, a functional language for programming network pipelines specifically targeted at multi-core scenarios. In addition to a large set of functions for generic packet manipulation, filtering, steering and state management, the framework is built upon an abstract model that provides state aware packet splitting to prevent inter-state sharing and enable consistent stateful parallel processing on-top-of multi-core architectures.

Mobile crowd sensing received significant attention in the recent years and has become a popular paradigm for sensing. It operates relying on the rich set of built-in sensors equipped in mobile devices, such as smartphones, tablets, and wearable devices. To be effective, mobile crowd sensing systems require a large number of users to contribute data. While several studies focus on developing efficient incentive mechanisms to foster user participation, data collection policies still require investigation. In this paper, we propose a novel distributed and sustainable framework for gathering information in cloud-based mobile crowd sensing systems with opportunistic reporting. The proposed framework minimizes cost of both sensing and reporting, while maximizing the utility of data collection and, as a result, the quality of contributed information. Analytical and simulation results provide performance evaluation for the proposed framework by providing a fine-grained analysis of the energy consumed. The simulations, performed in a real urban environment and with a large number of participants, aim at verifying the performance and scalability of the proposed approach on a large scale under different user arrival patterns.

In LTE and LTE-Advanced systems the rate of requests on the Random Access CHannel (RACH) can be high. Indeed, the Machine Type Communication (MTC) implies to have a high number of devices that need to request radio resources for transmitting small amount of data. Furthermore, reducing the time in which radio resources are allocated to Human Type Communications (HTC) for energy savings purposes, may lead to radio access network overload as well. In this framework, this paper aims at providing a set of guidelines for the resource allocation task in the RACH. In particular, the study investigates the impact of both the backoff indicator scheme and the maximum number of retransmissions on the RACH performance parameters. The rate of RACH requests associated with the HTC traffic is modelled by inferring their statistical properties starting from a dataset acquired in an operational eNodeB. The estimation of the average delay and the average number of maximum retransmissions gives insights on how many preambles should be reserved for HTC in order to meet the target performance, and provides suggestions on the configuration of the backoff indicator.

We present an SDN-based orchestration system that enables automated and coordinated arrangement of both computing and network resources in virtualized DC environments. The orchestrator is validated through a set of simulations, as well as through experiments carried out in a real laboratory. Results show that the proposed orchestration system improves the service requests acceptance rate and the utilization of network resources, while minimizing degradations of the user experience. Thus, DC resources can be effectively exploited, and different management objectives (e.g., load balancing, energy saving) can be reached.

This paper presents Packet Family Queue (PFQ), a high-performance framework for packet processing designed to flexibly handle network applications parallelism and making traffic processing safe and easy. PFQ is an open-source module for the Linux kernel that combines software-accelerated packet I/O to in-kernel early stage packet processing and fine-grained distribution to network applications and physical devices. PFQ does not require any modification to network device drivers and exposes programming interfaces to multi-threaded applications natively designed to run on top of it, as well as to legacy monitoring tools using the pcap library. The results show that the flexibility and the backward compatibility provided by PFQ do not impact its processing performance that, in fact, reaches line rate figures in the cases of pure speed tests and real practical monitoring use cases on 10+ Gb/s links.

In the last years, Software Defined Networking has emerged as a promising paradigm to foster network innovation and address the issues coming from the ossification of the TCP/IP architecture. The clean separation between control and data plane, the definition of northbound and southbound interfaces are key features of the Software Defined Networking paradigm. Moreover, a centralised control plane allows network operators to deploy advanced control and management strategies. Effective traffic engineering and resources management policies allow to achieve a better utilisation of network resources and improve end-to-end service performance. This paper deals with the architectural design and experimental validation of a control application that enables differentiated routing for traffic flows belonging to different service classes. The new control application makes routing decisions leveraging on OpenFlow network statistics, i.e., taking advantage of real-time network status information. Moreover, a Deep Packet Inspection module has been developed and integrated in the control application to detect VoIP traffic with Session Initiation Protocol signalling, enforcing this way policies for a differentiated treatment of VoIP traffic. Finally, a functional validation is performed in emulated environment.

The constant growth in IT as well as the increase of users and services requirements are making communication networking more and more complex to manage. This is why there has been the need to look for new approaches. One very promising solution is Software Defined Networking (SDN) that decouples the data and control planes, having one centralized controller for the network. This gives chances to control and manage the network as desired, thus opening many new possibilities. In this work we chose to exploit SDN to enable QoS control, starting from existing tools (Floodlight) and making modifications in order to better exploit their potential. A new network control application for end-to-end QoS provisioning is built on top of Floodlight and validated in an emulated SDN environment.

In the LTE systems, battery lifetime and network traffic overhead on control plane may be largely affected by the Discontinuous reception (DRX) configuration and the Radio Resource Control (RRC) Inactivity Timer. In this scenario, the paper proposes an analysis aimed at defining how to properly set the RRC Inactivity Timer to achieve a trade-off between energy savings and traffic overhead on the control plane. The analysis is based on an energy consumption model, whose key parameters are inferred directly from passive measurements carried out by monitoring a commercial eNodeB of one of the Italian Mobile Operators. The results suggest that taking into account the network traffic characteristics is possible to find the RRC inactivity timer value that permits to save energy of user's device while the increase of the signalling load is limited.

Today's rapidly evolving network ecosystem, characterized by increasing traffic volumes, service heterogeneity and mutating cyber-threats, calls for new approaches to packet processing to address key issues such as scalability, flexibility, programmability and fast deployment. To this aim, this paper explores a new direction to packet processing by pushing forward functional programming principles in the definition of a ''software defined networking'' paradigm.

Any improvement in packet classification performance is crucial to ensure Internet functions continue to track the ever-increasing link capacities. Packet classification is the foundation of many Internet functions: from fundamental packet-forwarding to advanced features such as Quality of Service en-forcement, monitoring and security functions. This work proposes a novel trie-based classification algorithm, named Jump-Ahead Trie (JA-trie), utilizing an entropy-based pre-processing phase and a novel approach to wildcard matching. Through extensive experimental tests, we demonstrate that our proposed algorithm is able to outperform a range of state-of-the-art classification algorithms.

Nowadays, Data Centers are the primary infrastructures for Cloud Computing services provisioning. In this challenging scenario, Data Centers have to deal with highly dynamic workloads, thus not only should they adopt advanced software virtualization solutions, but also network control capabilities. Elastic, agile network connectivity control should be integrated with the computing management infrastructure to make it achieve its goals. This paper presents an OF-based controller, called OFVN, that enables novel virtualization-aware networking functions in Cloud DCs. Virtual Machines allocations are performed considering the availability of computational resources on physical server, but also selecting a forwarding path for VMs data flows based on network links utilization. The focus of the paper is on the implementation and assessment of the controller in an experimental testbed. The behaviour of different allocation strategies is also investigated.

Oversubscription of intra-Data Center network links and high volatility of VM deployments require a flexible and agile control of Data Center network infrastructures, also integrated with computing and storage resources. In this scenario, the Software-Defined Network paradigm and, specifically, the OpenFlow protocol, opens up new opportunities for the design of innovative resource management platforms that enable dynamic and fine-grain control of DC networks through traffic engineering algorithms. This paper investigates the performance of two different sets of cloud-fluent traffic engineering algorithms. Conceived to work during cloud service deployments, the main target of such algorithms is to achieve a better utilization of network resources by exploiting OpenFlow capabilities for traffic-aware deployments of Virtual Machines. The effectiveness of the proposed solutions is evaluated in terms of network link utilization against VM requests acceptance ratio through simulations and experimental tests carried out by using an ad-hoc emulator.

This demo deals with a new orchestrator, the OFVN controller, specifically designed, according to the SDN principles, to deliver Cloud services with IT and network resources requirements.

Server virtualization enables dynamic workload management in data centers. However, some aspects of virtualization software technologies, like Virtual Machines (VMs) migration or communications between VMs and storage resources, can lead to huge and unbalanced utilization of the intra-data center network. In this paper, we investigate such issues in an experimental testbed, focusing on the measurement of the traffic overhead due to VMs migration in different operating conditions. Our measurement campaigns highlight that performance is strongly affected by several factors, such as VMs placement, VMs active memory, usage of Jumbo frames instead of Ethernet frames. We also analyze and compare the traffic exchanged between hosts on which VMs are placed, when two different storage systems (i.e., NFS and iSCSI) are used.

The evolution of commodity hardware makes it a very attractive platform to develop high-performance networking applications that are affordable to deploy. All but the most trivial applications must copy packets into user-space for further analysis. Therefore, the allocation of memory for these copies becomes a performance-critical operation. In this work, we present a multi-layer slice memory allocator specifically designed to take advantage of spatial and temporal locality in dealing with high-speed packet processing applications. Experimental results show that the proposed approach clearly outperforms existing memory allocators in common networking use-cases.

The detection of anomalies in backbone networks is posing serious performance issues, not only in terms of accuracy, but also in terms of detection speed. Indeed current software solutions to the problem, even promising from the point of view of detection and false alarm rates, suffer from the inability of performing the required operations in real time, when working in high speed backbone networks. On the other hand, hardware solutions are based on costly and inflexible niche systems.

Today's e-Infrastructures are deployed to enable easy and secure access to heterogeneous resources (e.g., instrument, computing and storage elements) by exploiting cloud and grid technologies. Since e-Science applications are usually characterized by huge data transfers, user interactivity and high computational loads, the selection and allocation of cloud or grid resources dramatically affect their performance. Starting from a previous work by the same authors, this paper proposes a new distributed resource allocation algorithm, referred to as MRA3D-H, capable of handling multiple resource requirements for jobs/tasks that arrive to the computing environment of the e-Infrastructure. More specifically, MRA3D-H aims at finding a trade-off between execution time and cost of data-intensive applications by taking into account performance parameters at system and network levels as well as the economic cost of computational resources. Simulations have been carried out to compare the performance of MRA3D-H with other resource allocation algorithms in the DORII project environment.

In the last years CUSUM based algorithms have emerged as a good basis to develop efficient systems able to detect anomalies and attacks in the network traffic. Nevertheless, such techniques are still far from providing an ideal solution to the problem, mainly because of the huge number of false alarms that characterizes such approaches. For this reason in this paper we propose a novel detection method based on the combined use of the CUSUM algorithm and the wavelet analysis. The introduction of the wavelet analysis, as shown by the obtained results, allows us to strongly reduce the false alarm rate of the “classical” methods, still retaining excellent performance in the detection of network anomalies.

The increasing number of network attacks causes growing problems for network operators and users. Thus, detecting anomalous traffic is of primary interest in IP networks management and many detection techniques, able to promptly reveal and identify network attacks, mainly detecting Heavy Changes (HCs) in the network traffic, have been proposed. Nevertheless, the recent spread of coordinated attacks, that occur in multiple networks simultaneously, makes extremely difficult the detection, using isolated intrusion detection systems that only monitor a limited portion of the Internet. For this reason in this paper we propose a novel distributed architecture that represents a general framework for the detection of network anomalies. The performance analysis, presented in this paper, demonstrates the effectiveness of the proposed architecture.

Network measurement both as diagnostic and within measurement-based techniques of traffic engineering and management, alongside network measurement for security has maintained the needs of researchers and network operators for the ongoing development of measurement tools for traffic monitoring/characterisation and to support Intrusion Detection Systems (IDSs). Many such tools capitalise on the pricing of commodity hardware by operating on general purpose architectures. Many are based on the well known libpcap API, a de facto standard in this area. Despite the many improvements that have been applied to packet capturing, packet-monitoring implementations still suffer from either: performance flaws on commodity hardware due mainly to unresolvable hardware bottlenecks, or costly and inflexible niche systems. To address such issues, the paper proposes a system architecture based on the cooperation of NetFPGA and a general purpose host PC. The NetFPGA is an open networking platform accelerator that enables rapid development of hardware-accelerated packet processing applications. The objective is to combine the high performance of a hardware-oriented solution with the flexibility of general purpose PCs.

The increasing number of network attacks causes growing problems for network operators and users. Thus, detecting anomalous traffic is of primary interest in IP networks management and many detection techniques, able to promptly reveal and identify network attacks, mainly detecting Heavy Changes in the network traffic, have been proposed. Among these, one of the most promising approach is based on the use of the CUSUM (CUmulative SUM). Nonetheless, CUSUM performance is strongly affected by its sensitivity to the presence of seasonal trends in the considered data. For this reason, in this paper we propose a novel detection method based on the idea of performing a pre-processing stage of the data by means of wavelets, aimed at filtering out such trends, before applying the CUSUM algorithm. The performance analysis, presented in the paper, demonstrates the efficiency of the proposed method, focusing on the performance improvements due to the pre-processing stage.

The choice of the path computation algorithm is a key factor to design efficient traffic engineering strategies in multi-protocol label switching networks and different approaches have been proposed in the literature. The effectiveness of a path computation algorithm should be evaluated against its ability to optimise the utilisation of network resources as well as to satisfy both current and future label switched paths allocation requests. Although powerful and flexible simulation tools might be useful to assist a network manager in the selection of proper algorithms, state-of-the-art simulators and network planning tools do not currently offer a suitable support. This study deals with the design and performance evaluation of multi-constraints path computation algorithms. To this aim, ad hoc software modules have been developed and integrated within the MTENS simulator. New single-path and multi-path computation algorithms have been proposed and compared in terms of number of accepted requests, success probability, network resources utilisation and execution time. Finally, some guidelines and recommendations for the selection of path computation algorithms have also been provided.

In this paper we propose a novel hardware-software co-design vision that aims at enhancing flexibility and reusability of hardware based packet forwarding engines. In particular, we move on the path of the well-known OpenFlow architecture that allows the user to decide the action to be performed over the packet (drop, forward through a given port etc.) upon interaction with a software control plane. Although such an approach is certainly powerful and is gaining more and more attention in both academia and industry, it is biased towards routing application: its main goal is to allow the software control plane to arbitrarily route a packet flow. However, we think that a similar paradigm, encompassing high performance packet forwarding hardware driven by a flexible software control plane, may be beneficial even to other kinds of applications, like monitoring and measurements. However, the primitives that the OpenFlow protocol provides are not flexible enough for such purposes. For this reason, we propose a flexible packet forwarding architecture based on regular expression that, besides enabling standard-compliant OpenFlow switching, can be easily reconfigured through its control plane to support other kinds of applications.

Modern network devices need to perform deep packet inspection at high speed for security and application- specific services. For this purpose, regular expressions are used, due to their high expressive power, and Deterministic Finite Automata (DFAs) are adopted to match them. Many works have been proposed to improve DFAs, especially in terms of memory consumption and speed. Instead, we address another issue: the scalability of DFAs to parallel systems and their buffer requirements. To our knowledge, a single attempt to parallelize DFA walk on regular multicore systems (which ex- ploits speculation with limited efficiency) has been proposed in literature. We propose a solution in which a number of processing units are committed to walk in parallel a DFA for the same packet; at this aim, sampling techniques on both text and regular expressions are adopted. This scheme is the first in literature that proposes effective parallelization of DFA walk, hence allowing for packet processing time reduction and less memory for reordering buffers. The result is that speed scales as the number of processing units.

The paper describes and compares different approaches that can be used to design Network Power Management methods, with the aim of reducing the power consumption of telecommunication networks. The approaches are based on the solution of optimization problems that have, in general, a Mixed Integer NonLinear Programming (MINLP) formulation. Given that the problems are NP-hard, exact methods for finding optimal solutions can be used only for scenarios of limited size. In this framework, the paper proposes a heuristic for finding a suboptimal solution of the Power Aware Routing and Network Design (PARND) problem, which is one of the more general design problems in Network Power Management. The simulation study highlights the capability of the proposed heuristic to obtain solutions near the optimum and to outperform the other approaches in terms of energy savings, while satisfying the constraints of the traffic demands.

Wavelength Routed Photonic (or all-optical) Networks (WRPNs) with an intelligent control plane will play a key role in providing the proper ground for future Internet applications. However, classic Routing and Wavelength Assignment (RWA) solutions are inadequate to be adopted by WRPNs. The main issues regard physical impairments and QoT, processing time and setup delay, lightpath (LP) survivability and minimization of resource allocation. In case of critical services provisioning, LPs must have tight QoT guarantees, even in case of network instability and failures. This work presents two new impairment-aware RWA (IA-RWA) algorithms, specifically tailored for the provisioning of critical services. These algorithms guarantee absolute bandwidth and Bit Error Rate (BER) for LPs, assuring survivability in case of single network failures. A high parallelizable, combined RWA engine minimizes resource utilization without increasing setup time. The performance of the proposed algorithms are evaluated with respect to a modified version of a previously proposed IA-RWA algorithm that takes into account QoT and survivability.

The increasing number of network attacks causes growing problems for network operators and users. Thus, detecting anomalous traffic is of primary interest in IP networks management. In this paper we address the problem considering a method based on PCA for detecting network anomalies. In more detail, we present a new technique that extends the state of the art in PCA based anomaly detection. Indeed, by means of the Kullback-Leibler divergence we are able to obtain great improvements with respect to the performance of the "classical" approach. Moreover we also introduce a method for identifying the flows responsible for an anomaly detected at the aggregated level. The performance analysis, presented in this paper, demonstrates the effectiveness of the proposed method.

The increasing number of network attacks causes growing problems for network operators and users. Thus, detecting anomalous traffic is of primary interest in IP networks management. In this paper we present a novel method for network anomaly detection, based on the idea of discovering Heavy Change (HC) in the distribution of the Heavy Hitters in the network traffic. To assess the validity of the proposed method, we have performed an extensive experimental evaluation phase, during which our system performance have been compared to a more “classical” HC-based approach. The performance analysis, presented in this paper, demonstrates the effectiveness of the proposed method.

The increasing number of network attacks causes growing problems for network operators and users. Thus, detecting anomalous traffic is of primary interest in IP networks management. In this paper we present a novel method for network anomaly detection, based on the idea of discovering Heavy Change (HC) in the distribution of the Heavy Hitters in the network traffic. To assess the validity of the proposed method, we have performed an extensive experimental evaluation phase, during which our system performance have been compared to a more "classical" HC-based approach. The performance analysis, presented in this paper, demonstrates the effectiveness of the proposed method.

Recently, the power consumption of the home/access networks has stimulated several research efforts devoted to the study of new mechanisms to optimize the energy utilization. In particular, most of these works is based on the idea that Ethernet links are mostly underutilized, and is thus aimed at realizing mechanisms able to adapt device consumption to the link utilization. Given the few works presenting an experimental study of the current Ethernet network devices from an energy point of view, this paper aims at presenting an extensive measurements campaign for evaluating the power consumption of two COTS Ethernet NICs. The measurements campaign, carried out following the procedures provided by the ECR Initiative™ and the SNE Energy Star, shows the energy inefficiency of the current NICs. Finally, using the experimental results, the paper presents a qualitative estimation of the energy savings achievable by implementing a well-known energy savings mechanism (i.e., ALR) in the Ethernet NICs.

Many wireless technologies used to build local or personal area networks (WLANs, WPANs) operate into the 2.4 GHz ISM band. Due to the mutual interference, the coexistence of such devices working at the same time in the same area can be troublesome. This paper reports the result of an extensive experimental study on the performance of three popular technologies (ZigBee, Wi-Fi, and Bluetooth), under their mutual disturbance. In addition to assessing the effect of the interference between pairs of technologies, we also analyzed the simultaneous interference among the three systems, a scenario that has never been studied before. Our work partly confirms previous findings, but also reveals some unexpected behaviors, imposing to reconsider some assumptions about the supposed interference-free ZigBee and Bluetooth channels.

Overload is a challenging problem for a SIP server because the built-in overload control mechanism based on generating rejection messages could not prevent the server from collapsing due to congestion. In this scenario, the paper presents an overload mechanism combining a local and a remote solution. The local part of the overload control mechanism is based on the appropriate queueing structure and buffer management of the SIP proxy. The remote overload control mechanism is based on feedback reports provided by the SIP proxy to the upstream neighbors. These reports permit the traffic regulation necessary to avoid the critical condition of overload. The main paper contributions are the design of key components of a remote control mechanism, the proposal of a new approach for dynamic load estimation, and the use of a prediction technique in the remote control loop.

Deep packet inspection is a fundamental task to improve network security and provide application-specific services. State-of-the-art systems adopt regular expressions due to their high expressive power. They are typically matched through deterministic finite automata (DFAs), but large rule sets need a memory amount that turns out to be too large for practical implementation. Many recent works have proposed improvements to address this issue, but they increase the number of transitions (and then of memory accesses) per character. This paper presents a new representation for DFAs, orthogonal to most of the previous solutions, called delta finite automata ($\delta$FA), which considerably reduces states and transitions while preserving a transition per character only, thus allowing fast matching. A further optimization exploits$N$th order relationships within the DFA by adopting the concept of “temporary transitions.”

An e-Infrastructure allows end-user's applications to easily and securely access heterogeneous grid resources (e.g., computing and storage elements). Since e-Science applications are often characterized by huge data transfers and high computational loads, the selection and allocation of grid resources dramatically affect their performance. This paper proposes a distributed resource allocation algorithm, referred to as MRA3D, capable of handling multiple resource requirements for jobs/tasks submitted to the grid computing environment of the e-Infrastructure. More specifically, MRA3D aims at minimizing the execution time of data-intensive applications by taking into account performance metrics describing both system and connectivity status of the computational resources. Simulations have been carried out to compare the performance of MRA3D with other resource allocation algorithms in a realistic environment by using synthetic as well as real workload traces.

DiffServ-aware MPLS-TE (DS-TE) significantly enhances the MPLS-TE architecture by enabling bandwidth reservation on a per-class basis. More specifically, to this aim, the so-called Bandwidth Constraint Models (BCMs) are introduced as a key architectural element. This paper describes a new NS2 module that offers a complete support for the DS-TE architecture including the standard BCMs (MAM and RDM) and a new BCM (G-RDM), recently proposed by the authors. Moreover, the paper shows the results of simulations carried out to evaluate the behaviour of the above mentioned BCMs in a real network scenario.

Because of the rapid growth of both traffic and links capacity, the time budget to perform IP address lookup on a packet continues to decrease and lookup tables of routers unceasingly grow. Therefore, new lookup algorithms and new hardware platform are required to perform fast IP lookup. This paper presents a new scheme on top of the NetFPGA board which takes advantage of parallel queries made on perfect hash functions. Such functions are built by using a very compact and fast data structure called Blooming Trees, thus allowing the vast majority of memory accesses to involve small and fast on-chip memories only.

A Bloom Filter is an efficient randomized data structure for membership queries on a set with a certain known false positive probability. Bloom Filters (BFs) are very attractive for their limited memory requirements and their easy construction which make them a popular choice for many tasks in network devices. However, in a number of network applications, more than simple probabilistic membership queries is required, and BFs can be adopted as a coarse filtering stage, leaving the ultimate filtering and classification process to other techniques, such as hash tables or tree-like structures. In this paper we propose a scheme to extend BFs with "indexing" features so that when an element x is queried, an univocal index of that element is returned, which in turn can be used as an address for a table, just as a perfect hashing scheme. This extension, called indexed Bloom Filter (iBF), comes at the cost of a small increment of false positive probability and simply fits in existing BF-based applications.

Modern network devices need to perform deep packet inspection at high speed for security and application-specific services. Instead of standard strings to represent the dataset to be matched, state-of-the-art systems adopt regular expressions, due to their high expressive power. The current trend is to use Deterministic Finite Automata (DFAs) to match regular expressions. However, while the problem of the large memory consumption of DFAs has been solved in many different ways, only a few works have focused on increasing the lookup speed. This paper introduces a novel yet simple idea to accelerate DFAs for security applications: payload sampling. Our approach allows to skip a large portion of the text, thus processing less bytes. The price to pay is a slight number of false alarms which require a confirmation stage. Therefore, we propose a double-stage matching scheme providing two new different automata. Results show a significant speed-up in regular traffic processing, thus confirming the effectiveness of the approach.

In the last few years, the number and impact of security attacks over the Internet have been continuously increasing. Since it seems impossible to guarantee complete protection to a system by means of the "classical" prevention mechanisms, the use of Intrusion Detection Systems has emerged as a key element in network security.

Bloom Filters are efficient randomized data structures for membership queries on a set with a certain known false positive probability. Counting Bloom Filters (CBFs) allow the same operation on dynamic sets that can be updated via insertions and deletions with larger memory requirements. This paper first presents a simple tight upper bound for counters overflow probability in CBFs, which is adopted in the design of more efficient CBFs. On the basis of such theoretical achievements, we introduce the idea of a hierarchical structure as well as the use of Huffman code to improve standard CBFs in terms of fast access and limited memory consumption (up to 50% of memory saving). The target could be the implementation of the compressed data structures in the small (but fast) local memory or “on-chip SRAM” of devices such as network processors. As an application of our algorithms, an anti-evasion system is finally proposed.

Exact and approximate membership lookups are among the most widely used primitives in a number of network applications. Hash tables are commonly used to implement these primitive functions as they provide O(1) operations at moderate load (table occupancy). However, at high load, collisions become prevalent in the table, which makes lookup highly non-deterministic and reduces the average performance. Slow and non-deterministic lookups are detrimental to the performance and scalability of modern platforms such as ASIC/FPGA and multi-core that use highly parallel compute and memory structures.

IP address lookup is a fundamental task for Internet routers, due to the rapid growth of both traffic and links capacity. Many algorithms have been proposed to improve lookup performance in terms of memory consumption, search speed and update complexity. Due to the presence of wildcards and netmasks, such algorithms adopt several techniques to deal with longest prefix matching. However, the analysis of lookup tables reveals that the first 16 bits of forwarding rules are almost always specified. Therefore, more powerful exact-matching schemes can be applied to the first half of addresses. This paper presents a routing lookup accelerator (RLA) which allows the lookup of the first 16 bits to be sped up. The target is an efficient scheme to be implemented in small and fast memories of recent hardware platforms. Specifically, since in several forwarding tables the distribution of the first 16 bits is characterized by empty gaps as well as pronounced peaks, we propose to divide the address space in different ranges and to encode each address only as a difference with respect to a given address chosen as reference for that range. Then, a hybrid direct-addressing / multibit trie scheme is used for each range. As RLA is orthogonal to all other schemes, any other lookup algorithm can be used to perform longest prefix matching on the remaining bits.

Characterizing delay distribution over the links of a network provides a remarkable amount of information which can be useful for troubleshooting, traffic engineering, adaptive multimedia flow coding, overlay network design, etc. Since querying each and every node of a path in order to retrieve this kind of information can be unfeasible or just too resource demanding, the recent research trend is to infer the internal state of a network by means of end-to-end measurements. Many algorithms in literature require active measurements and are based on a single-sender multiple-receivers scheme, thus relying on the cooperation of a possibly wide number of nodes, which is a quite strong assumption. Moreover, many previous works adopt Expectation-Maximization algorithms to cope with large and under-determined equation systems, thus increasing the uncertainty of the final delay estimation. This paper, instead, proposes a technique to infer the cumulants of the delay distribution over each link of a given network path, based on two-points measurements only. The cumulants, in turn, can be used to approximate the distribution function through the Edgeworth series. The results of our approach are assessed through a wide series of model-based and ns2 based simulations and show fairly good performance under different network load conditions.

Network Topology Discovery is crucial to a number of network management tasks. Traditional topology discovery techniques require internal nodes to take actions on measurement packets, which makes them unpractical in many cases. For these reasons, tomographic techniques have been introduced, which allow for the reconstruction of network topologies with no need for cooperation from internal routers. The usual approach to tomographic topology discovery is based on clustering nodes into tree structures according to soft similarity metrics. We recently proposed a novel technique based on decision theoretic considerations that help the topology reconstruction by limiting the set of hypotheses to a finite and well-defined set, thus determining hard metrics. In the scheme, probe traffic is sent to all couples of end-nodes and a metric is assigned to each measurement. In this paper, we extend the technique by ordering the topology reconstruction procedure according to metrics reliability defined in terms of their variances. The algorithms presented in the paper are validated through extensive simulations in several network scenarios. The results show that such a methodology allows to retrieve a complete picture of the network that includes the detection of all the internal nodes along with the values of capacities of the interconnecting links.

Deep packet inspection is required in an increasing number of network devices, in order to improve network security and provide application-specific services. Instead of standard strings to represent the data set to be matched, state-of-the-art systems adopt regular expressions, due to their high expressive power and flexibility. Typically regular expressions are matched through deterministic finite automata (DFAs), but large rule sets need a memory amount which turns out to be too large for practical implementation. Many recent works have proposed improvements to address this issue, but they increase the number of transitions (and then of memory accesses) per character. In a previous work, we have presented a smart representation for DFA which, while preserving fast matching (i.e., a transition per character only), considerably reduces states and transitions. In this paper we introduce a novel optimized automaton, which exploits second order relationships within the DFA and is based on the key concept of "temporary transitions". Results for real data sets show that it allows for a further memory saving.

The paper presents a scheduling algorithm for WiMAX systems, which exploits Time-Utility Functions (TUFs) and game theory. The scheduling decision is performed by means of two steps: the intra-class scheduling step and the inter-class scheduling step. In the first one, each user exploits TUFs to choice the packets to transmit according to QoS requirements. In the second one, the scheduler exploits the game theory to select a specific game solution that provides more or less fairness among the users that have to be served. The available physical network resources (i.e. slots) are taken into account as a constraint to the scheduling decision. The proposed scheduling algorithm follows a cross-layer approach that takes into account application requirements by means of TUFs and physical Channel State Information (CSI) of each user. Several game solutions are compared by means of simulation results in terms of user fairness, user throughput, packet delay and packet expiration rate. The TUFs introduction is also evaluated comparing the different applications performance.

As other signaling protocols in the past, also SIP suffers of server overload leading to performance collapse. In this framework, recent Internet drafts propose the improvements of overload control mechanisms already presented in SIP and/or a closed loop system model for overload avoidance. In this framework, the paper presents the system simulator developed extending the network simulator (ns-2) with logical for local overload control. Furthermore, the paper proposes a new queueing discipline obtained combining the simple first in first out (FIFO) service discipline with the priority one. Then the paper presents a simulation analysis, aimed at evaluating the impact on system performance of different queueing structures, service disciplines, and buffer sizes. Simulation results clearly show that the proposed queueing discipline produces good system performance with a low complexity increase. Finally, the simulation results point out the weakness of the 503 service unavailable message mechanism, which does not introduce significant improvement when combined with the proposed solution.

The paper reports an analysis about the convenience of employing multi-protocol label switching (MPLS) to provide mobility in wireless mesh networks (WMNs). In such networks, which are expected to serve a considerable number of mobile users, the efficient support of mobility is a non-trivial issue. The starting point of our work is proxy mobile IPv6 (PMIPv6), a standardised network-based protocol for localised mobility management that has already shown good performance. We analyse and compare PMIPv6 with a recently proposed extension that aims at employing MPLS tunnels in place of IPv6 tunnels. Our contribution is novel in two aspects. First, the analysis focuses on a recent application scenario (the WMN) and accounts for its typical features. Second, we investigate a non-standard use of MPLS (i.e. for micro-mobility management) and provide the first evaluation work of the MPLS tunnel for PMIPv6 technique.

Nowadays, many IP backbone networks adopt separate control and forwarding planes for unicast and multicast traffic flows. Indeed, while MPLS (multiprotocol label switching) is widely deployed for unicast traffic, IP multicast is the only available solution for the delivery of "one-to-many" traffic flows. With P2MP LSPs (point-to-multipoint label switched paths) support, a unified control and forwarding plane may be devised. Such a reduction in the number of protocols used in the core of the network as well as in the number of encapsulations in the data plane, results in simplified network operations. The paper discusses the design and the development of the control and data planes extensions needed to provide P2MP LSP support in an MPLS node. In particular, such extensions, concerning the P2MP LSPs path computation, the RSVP-TE signalling protocol, and the forwarding mechanism, have been implemented as new software modules for an ad-hoc developed simulator, based on NS2. Finally, tests have been performed to assess the behaviour of the new functionalities introduced in the simulator.

In the last years, broadband satellite networks have emerged as a flexible technology offering high-speed Internet access to large communities of users at relatively low-cost. At the same time, the large widespread of VoIP and Multimedia over IP (MoIP) applications, has pointed out the need of investigating their performance over heterogeneous networks and, more specifically, on wireless and satellite systems not offering Quality of Service (QoS) guarantee. The paper presents the results of an experimental study aimed at analyzing the behaviour of MoIP applications over a two-way satellite system. Starting from the most relevant performance metrics of a satellite network, the paper is focused on the evaluation of the Perceived Quality of Service (PQoS) when VoIP and MoIP applications are deployed over a satellite network. Control and data plane issues that may deteriorate the performance of such applications are highlighted.

In the last few years, the number and impact of security attacks over the Internet have been continuously increasing. Since it seems impossible to guarantee complete protection to a system by means of the "classical" prevention mechanisms, the use of Intrusion Detection Systems has emerged as a key element in network security. In this paper we address the problem considering some techniques for detecting network anomalies. Our approach is based on the use of different compression algorithms for detecting anomalies in the network traffic running over TCP. In more detail we take into account the use of three different compression algorithms, based on distinct approaches, namely: Huffman coding, Dynamic Markov Coding, and Lempel-Ziv-Welch algorithm. The proposed methods are based on the consideration that the entropy represents a lower bound to the compression rate that we can obtain, and that the more redundant the data are and the better we can compress them. The performance analysis, presented in this paper, demonstrates the effectiveness of the proposed methods.

The paper presents a measurement campaign carried out on a 2.5 GHz Mobile WiMAX test-bed in a sea port scenario, aimed at evaluating the system performance. The analysis of acquired data suggested the definition of a 2-ray pathloss model, which is able to fit the actual behavior of the observed marine channel. In order to evaluate the effects of the proposed pathloss model on system performance the paper presents a simulation study carried out by means of OPNET simulator. In particular, the simulation study is aimed at comparing the throughput at IP level obtained using different pathloss models implemented in OPNET, with that obtained when the proposed 2-ray model is considered. Further, the throughput at IP level obtained by simulation is compared with the one measured in the actual test-bed. This comparison points out that the WiMAX OPNET model is able to reproduce in most cases the system performance observed during the measurement campaign, when the proposed 2-ray pathloss model is used. However, the simulator is not able to predict the behavior of actual system in all situations. Indeed, in the areas where the pathloss presents peaks, the breakdown of system performance, mainly caused by synchronization problems, is not predicted by the simulator.

In the last years, grid computing has emerged as a valuable service to solve complex computational problems in many scientific and industrial domains. Quality of Service (QoS) provision for these applications is therefore a key challenge for high speed Next Generation Networks and cross-layer mechanisms, enabling the development of network-aware grids, should be introduced. This paper takes into account, as a case study, an application for the simulation of particles coagulation phenomena. A parallel direct simulation Monte Carlo algorithm has been implemented and extensive computations have been performed in a cluster of computing nodes to evaluate the execution time of the application in different operating conditions. Then, data collected from these experiments, have been used as input for the design and deployment of a network-aware grid infrastructure. Since this application is characterized by frequent data exchanges, the impact of the allocated bandwidth on the overall application's performance has been investigated.

Tomographic techniques allow the reconstruction of network topologies with no need for cooperation from internal routers. However, most of such mechanisms adopt a method of node clustering producing trees that reveal only a partial structure of the network. Therefore, we have proposed a novel approach to topology discovery based on packet sandwich probes and decision theory allowing to retrieve a complete picture of the network, which includes the detection of all the internal nodes along with the values of capacities of the interconnecting links. Such an approach, as well as all the standard techniques of topology discovery, reconstructs the spanning tree of the probe sender only. Hence, in this paper a specific technique is presented for merging the spanning trees associated to all different roots, in order to provide a complete representation of the network. Such a method does not require further probing traffic and is specifically designed to merge topology reconstructions where all the nodes of the network (not only the branching nodes) are revealed, along with link capacities. Our algorithm performs quite well on a wide set of both synthetic and realistic topologies, and in many cases provides a picture of the network which is exactly equivalent to the original one.

Performing deep packet inspection at high speed is a fundamental task for network security and application-specific services. In state-of-the-art systems, sets of signatures to be searched are described by regular expressions, and finite automata (FAs) are employed for the search. In particular, deterministic FAs (DFAs) need a large amount of memory to represent current sets, therefore the target of many works has been the reduction of memory footprint of DFAs. This paper, instead, focuses on speed multiplication by enlarging the amount of bytes observed in the text (i.e., searching for k-bytes per state-traversal). For this purpose, an interesting yet simple inverse homomorphism is employed to reduce the amount of transitions in the modified DFA. The algorithm results to be remarkably faster than standard DFAs, and provides also a good compression scheme that is orthogonal to other schemes.

In the last few years the number and impact of security attacks over the Internet have been continuously increasing. Since it is impossible to guarantee complete protection to a system by means of the "classical" prevention mechanisms, the use of Intrusion Detection Systems (IDSs) has emerged as a key element in network security. In this paper we address the problem considering some techniques for detecting network anomalies, based on the use of co-occurrence matrices, to model the "normal" behavior of the TCP connections.

We propose a traffic engineering (TE) architecture to supply wireless mesh networks (WMNs) with an efficient support of diverse applications with different quality of service (QoS) requirements. The proposed TE-WMN architecture is based on the MPLS technology. Its main goal is to provide TE capabilities so that a heterogeneous infrastructure, made of wired networks and WMNs, can be built and managed in a seamless manner. In the TE-WMN architecture, the path computation element (PCE) is in charge of finding a path that satisfies the QoS requirements. This operation often requires solving a multi-constrained problem, which is known to be NP-complete. So far many approximated and heuristic solutions have been developed for and applied to wired networks. However, the different features of WMNs make it impractical to reuse those solutions. Hence we have devised two heuristic path computation algorithms (PCAs) with the goal of keeping the complexity very low and yet improving the performance of simple existing algorithms. An extensive set of tests has proved the good performance of the proposed PCAs.

The paper presents a network scenario for multimedia communications during emergency operations in a sea area. The network scenario is composed by a satellite link and a WiMAX system provided by a main naval unit. The WiMAX system permits the communications between the operators on small units and the main naval unit, where there are both local coordinators and the shared satellite link, used for communications between units on the crisis area and the headquarter. The presented network scenario has been experimentally tested in terms of coverage area offered by the WiMAX system, IP throughput and user level quality of service of VoIP applications.

Abstract Computer security is a growing problem. Over the last years, the number and variety of security attacks in IP‐ based network infrastructures have increasingly grown, leading to the need of developing new security architectures. In this scenario, the use of intrusion detection systems (IDSs) has emerged as a key element, since it permits to tackle security threats by masquerader, misfeasor, and clandestine users. In this paper, we address the problem considering some new statistical techniques for detecting network anomalies. In more detail, the paper discusses the use of several statistical models to characterize the normal behavior of the network traffic running over TCP, so that anomalies can be revealed as significant deviations from such behavior. Namely, our proposal is based on the use of Markov chains, co‐ occurrence matrices, and compression algorithms, for modeling the TCP connections, in terms of statistical analysis of some of the packet header fields. The performance analysis, presented in this paper, demonstrates the effectiveness of the proposed methods. Copyright © 2009 John Wiley & Sons, Ltd.

Evaluating the performance of high-speed networks is a critical task due to the lack of reliable tools to generate traffic workloads at high rates. The current open-source software tools are not suitable to deal with high-speed networks as they present poor performance in terms of generated frames per second and scarce timing/rate accuracy in traffic generation. These issues are due to the intrinsic limitations of the PC architecture, for which these tools are designed. This paper proposes a different approach based on the Intel Network Processor IXP2400. The design aims to maintain the high flexibility of PC solutions while outperforming them in terms of throughput and packet rate. This is obtained by combining a general-purpose PC with the processing units of a network processor.

Hash tables are used in many networking applications, such as lookup and packet classification. But the issue of collisions resolution makes their use slow and not suitable for fast operations. Therefore, perfect hash functions have been introduced to make the hashing mechanism more efficient. In particular, a minimal perfect hash function is a function that maps a set of n keys into a set of n integer numbers without collisions. In literature, there are many schemes to construct a minimal perfect hash function, either based on mathematical properties of polynomials or on graph theory. This paper proposes a new scheme which shows remarkable results in terms of space consumption and processing speed. It is based on an alternative to Bloom Filters and requires about 4 bits per key and 12.8 seconds to construct a MPHF with 3.8times109elements.

Tomographic techniques allow for the reconstruction of network topologies with no need for cooperation from internal routers. Traditional tomographic techniques infer the internal network layout by clustering nodes into tree structures that, in many cases, reveal only a partial graph structure of the network. This paper proposes a novel approach to network topology discovery by means of packet sandwich probes; the underlying theoretical basis relies on the application of Decision Theory to a finite set of possible topological hypotheses. The decision process is however disturbed by the interaction of probes with regular cross traffic, which results in a background noise that afflicts the measurements. To cope with this phenomenon, a model-free noise reduction technique is also used. The algorithms presented in the paper are validated through extensive simulations in several network scenarios. The results show that such a methodology allows to retrieve a complete picture of the network that includes the detection of all the internal nodes along with the values of capacities of the interconnecting links.

Originally designed to improve the efficiency of packets forwarding, MPLS provides support for Traffic Engineering and network resilience. Constrained-based path computation is a key building block for Traffic Engineering in MPLS networks, since it allows to set-up LSPs along paths that satisfy QoS constraints. This paper deals with on-line and offline path computation algorithms and, more specifically, introduces and compares the performance of three novel offline path computation algorithms which aim at improving the performance of their standard version by means of some heuristics. All these algorithms have been developed in NS2 as an extension of OSPF-TE\ns and integrated with RSVP-TE\ns.

A Bloom filter is an efficient randomized data structure for membership queries on a set with a certain known false positive probability. A counting Bloom filter (CBF) allows the same operations on dynamical sets that can be updated via insertions and deletions with larger memory requirements. This paper presents a novel hierarchical data structure, called Blooming tree, that replicates the functionalities of a CBF with lower memory consumption and tunable false positive probability. The hierarchical multi-layer design of Blooming trees allows for distributing the structure in different memory levels, thus exploiting small but fast on-chip memories for most frequently accessed substructures. The proposed algorithm is compared to previous existing schemes on a target platform: Intel IXP2XXX Network Processors (NPs).

The paper presents PingPair, a novel tool for end-to-end path capacity estimation. The tool is based on the classical packet dispersion technique, enhanced by a novel algorithm for the selection of the best measurement samples based on queueing delay estimation. In addition, PingPair takes into account the measurement noise that afflicts the interarrival times registered by a user level application; we experimentally observe the Gaussian nature of such a noise. Since PingPair relies on one- point measurements only, it can be deployed in almost all network scenarios, thus providing maximum flexibility. The performance of the tool has been assessed through both NS2 simulations and extensive experimental campaigns, including Internet as well as field trial measurements. The results are compared to those achieved by Capprobe, which is one of the most effective out of the many available one-point measurement-based capacity estimation tools. Despite the very low amount of probing traffic generated, PingPair outperforms Capprobe in most scenarios, yielding more precise capacity estimates; therefore, it proves to be a very fast and unintrusive way to measure the capacity of a network path.

Bloom filters are efficient randomized data structures for membership queries on a set with a certain known false positive probability. Counting bloom filters (CBFs) allow the same operation on dynamic sets that can be updated via insertions and deletions with larger memory requirements. This paper first presents a new upper bound for counters overflow probability in CBFs. This bound is much tighter than that usually adopted in literature and it allows for designing more efficient CBFs. Three novel data structures are proposed, which introduce the idea of a hierarchical structure as well as the use of Huffman code. Our algorithms improve standard CBFs in terms of fast access and limited memory consumption (up to 50% of memory saving): the target could be the implementation of the compressed data structures in the small (but fast) local memory or "on-chip SRAM" of devices such as network processors .

Global Grid Computing goal is to connect heterogeneous computational resources belonging to the same Virtual Organization (VO) through the Internet to form a single, more powerful virtual computer. However, to fulfill this goal it is necessary to develop services that provide the virtual computer with the same functionalities of individual end systems, such as security, interprocess communication, and resource management.

MPLS traffic engineering (TE) allows the creation of end-to-end paths across the network with bandwidth reservations. The main drawback of the basic MPLS-TE model is that it operates at an aggregate level and so it is unaware of traffic classes. DiffServ-aware MPLS-TE (DS- TE) refines the MPLS-TE model by allowing bandwidth reservations to be carried out on a per-class basis. The result is the ability to give strict QoS guarantees while optimizing the use of network resources. Bandwidth constraints models play a key role in the DS-TE architecture, since they establish how bandwidth is distributed among different classes. In this paper, we first present a new bandwidth constraints model, called G-RDM. Then, we compare the performance of G-RDM with respect to MAM and RDM in different scenarios by means of an analytical model based on Markov-chains. The results show that G-RDM, joining the best features of MAM and RDM, allows to improve their performance.

The growth of the Internet in the last years has been pushed by increasing requirements in terms of capacity, security and reliability. Moreover, improvements in multimedia applications need mechanisms and architectures to accomplish quality of service (QoS) and differentiated services. Technology development has shown that the evolution of processing power cannot cope with the link capacity growth. Therefore a link capacity scheduler is no longer sufficient to assure efficient service differentiation to end-users, but a proper computing power allocation for packet processing must be adopted. In this paper a processing scheduling scheme for IntelregIXP2XXX Network Processors is proposed. A model of the architecture is defined and an ad-hoc simulator is developed to help the comprehension of the system and the re-design of the application. Finally experimental results show the performance of the proposed algorithm.

In grid computing the need for collecting information about both distributed computational resources and network topology and performance is constantly growing. Several tools are currently under development to provide such information. They are based either on a centralized or a distributed architecture. Distributed tools are commonly based on IP-level application- oriented network metrology. The measurements are done by means of beacons running in some network nodes (e.g., grid hosts) and collecting the required information. However, the number of utilized beacons might heavily impact the final result, i.e. the discovered topology and the collected performance information. In this study a model is developed to estimate the percentage of discovered links provided that a specific number of beacons is placed in the network. The model is developed for Erdos-Renyi (ER) graph network models but it can be applied also to other networks. Numerical evaluation shows that the model closely approximate the percentage of discovered links obtained through simulation for ER networks. For other theoretical and real networks the model overestimates the percentage of discovered links. However experimental results show that for networks with realistic average nodal degree the overestimate is less than 20%.

The high level of performance achieved by todays traditional PCs makes x86 personal computers effective and inexpensive platforms for the development of high performance network devices. In particular, such devices can replace, especially in the access portion of the network, traditional devices for operations like management, monitoring and accounting. This new possibility pushes the research towards the development of new algorithms and functionalities on top of the x86 architecture. In this framework, the purpose of this work is to design a new high-speed longest prefix classification algorithm for packet accounting to be integrated on top of Linux based PCs for traffic characterization and accounting on high speed (> 1Gbps) links, regardless of the detail level of the rule set.

Originally conceived as a fast forwarding technique, MPLS provides support for traffic engineering and network survivability. Constrained-based path computation is a key building block for traffic engineering in MPLS networks, since it allows to select a path that satisfies assigned QoS requirements. In this paper, we introduce a novel path computation procedure which aims at improving the performance of the well-known Wang-Crowcroft algorithm by means of some heuristics. Moreover, the two algorithms have been developed in NS2 as an extension of OSPF-TE\ns and integrated with RSVP-TE\ns. Finally, the paper shows how the developed software module can be used to satisfy a set of LSP allocation requests with multiple QoS constraints.

The recently approved IEEE 802.11e standard improves the support for real-time multimedia applications in Wireless LANs. Yet, to provide some level of guarantee in terms of QoS parameters, it is essential to use an admission control scheme. We have devised one such scheme, which bases the admissibility test on the time occupancy of the medium. The network state is assessed through an analytical model of the EDCA in non-saturation conditions. In the paper, beyond describing the proposed scheme, we also show its effectiveness in the case of voice over IP applications.

In this paper we describe a time-based admission control scheme for multimedia over IP services in wireless LANs based on the IEEE 802.11e standard. We build on an existing work where it was defined a parameter, called network utilization characteristic (NUC), to measure the temporal occupancy of the wireless channel of an IEEE 802.11 network. The work we present is twofold. At first, we extend this parameter to 802.11e systems. A simple Markov chain is used for this purpose. Then we build an admission control scheme based on the so-extended NUC and verify its effectiveness in a practical case: a network with mixed voice and video traffic. Simulations, also with different link qualities, are used to validate the effectiveness of our scheme.

Topology discovery techniques based on a network tomography approach can be successfully adopted in almost all scenarios, in that they infer the internal characteristics of a network without any cooperation from the internal nodes. Out of the many tomographic topology discovery techniques proposed in the literature, those based on the use of packet sandwich probes (a special kind of packet trains) present some particularly attractive features. The rationale of such approaches is to take advantage of end-to-end measurements to infer the logical topology of the network through hierarchical clustering algorithms. Typically, due to the interference with cross traffic, such measurements are affected by a zero-mean noise which, in turn, may cause the wrong reconstruction of the network topology. This paper analyzes the causes of certain noise patterns (which have actually been observed during experiments) and proposes a noise reduction algorithm to sort out this issue. Such an algorithm does not rely on any assumption about the statistical model of the cross-traffic noise and its effectiveness has been tested through a campaign of ns2 simulations.

In previous works we have introduced the DTT (Deficit Transmission Time) scheduler, an algorithm to be run at the Access Point to contrast the performance anomaly affecting IEEE 802.11 based networks. DTT has indeed proved to be very effective, as it is able to guarantee proportional fairness and isolation to downlink flows. Yet, due to its centralized nature, it can only put in place a suboptimal solution, as the uplink flows are still uncontrolled We have therefore extended the idea of DTT to a distributed version of the scheduler, called Distributed DTT, or (DT)2. The paper, beyond describing this new algorithm, presents the results of an extensive set of simulations. These show that (DT)2is able to defeat the performance anomaly for uplink flows, thus complementing its ancestor perfectly. A useful side effect is the reduction of the bottleneck effect, that allows more fairness between uplink and downlink directions. We show that these achievements are crucial in supporting the emerging Voice Over WLAN services.

In a multi-service IP network, it is a key challenge to provide QoS to end-user applications while effectively using network resources. Traffic Engineering support in MPLS networks allows operators to carry traffic with strict QoS requirements by means of advanced network capabilities, such as resource reservation, fault-tolerance and optimization of transmission resources. To deliver a reliable service, MPLS exploits a set of procedures that protect the traffic flows carried on different paths. Therefore, Label Switching Routers have to support mechanisms for fault detection, notification and recovery, whereas MPLS has to enable the configuration of different recovery techniques. In this paper, we present the design and the development of some software modules for the simulation of different MPLS-based recovery strategies in the discrete event simulator NS2. Moreover, the paper discusses some simulation results that show a performance comparison between path protection and path restoration strategies in a simulation scenario matching the Geant network.

In operational networks, services are delivered according to specific service level agreements (SLAs). To ensure the SLAs stipulated with the users, providers make use of traffic monitoring and measurement tools controlling their highspeed IP networks. In order to improve the scalability of the measurement systems, techniques which allow to reduce the total amount of data to be collected and processed are currently being standardized in the IETF packet sampling (PSAMP) Working Group. Among packet sampling techniques, hash-based one is the only capable of providing a feasible framework to monitor one-way delays (OWDs). This paper is focused on the experimental validation of SLAs conformance in operational networks using hash-based packet sampling techniques. The results presented hereafter are obtained from the analysis of traffic traces collected at two different points of operational networks. Such results are fundamental to validate the whole hash-based packet sampling framework, because they complement the already available results on the uniformity of distribution, computation speed and statistical properties of the hash functions calculated by using single point traces.

In the paper we introduce and evaluate the performance of the Deficit Transmission Time (DTT) scheduler, a centralized channel aware scheduling algorithm for IEEE 802.11 networks. The scheduler has been designed to face two inconveniences of 802.11: the performance anomaly and the "inter-flow blocking" problem. DTT splits outgoing frames into multiple queues and uses gross transmission times as a measure of the link quality. DTT guarantees each downlink flow an equal time share of the channel, which translates into the desirable properties of proportional fairness and flow isolation. The performance of the scheduler has been measured in terms of the number of VoIP calls that can coexist in the system with an acceptable perceived speech quality. This assessment has been done on the basis of the E-model, an ITU-T standardized computational tool. The results definitely point out the improvement over the plain FIFO policy.

Nowadays packet classification is a fundamental task for network devices such as edge routers, firewalls and intrusion detection systems. Determining which flow packets belong to is important for many applications, and it is necessary, for example, to provide differentiated services, to detect anomalous traffic and to sort attack patterns. Therefore packet classification is becoming more and more complex, with more flexibility and higher performance requirements. Network Processors (NPs) are emerging as very promising platforms due to their capability to combine the flexibility of general-purpose processors with high performance of hardware-based solutions. In this paper we illustrate the design of a multidimensional packet classifier realized on the Radisys® ENP-2611 board equipped with Intel® IXP2400 Network Processor. The first goal of this study is the selection of the most suitable classification algorithm to be integrated into the embedded system. Our investigation is then directed to adjustments and refinements of the selected algorithm (namely the multidimensional multibit trie algorithm) to capitalize the peculiar functional properties and capabilities of our network processor.

In a multi-service IP network, it is a key challenge to provide quality of service (QoS) to end-user applications while effectively using network resources. The DiffServ architecture has emerged as a scalable solution to provide a network with multiple service classes. The multiprotocol label switching (MPLS) traffic engineering (TE), instead, enables resource reservation, fault-tolerance and optimization of transmission resources. MPLS DiffServ-aware TE combines the advantages of these solutions and allows network operators to provide services that require strict QoS performance guarantees. The choice of the signalling protocol to exchange DiffServ, MPLS and TE-related information is a key point in this scenario. Different solutions have been proposed, such as RSVP-TE and CR-LDP. Since the first one seems to be the most attractive for router manufacturers, in this paper, the design and the development of a new software module to simulate the RSVP-TE protocol in the Network Simulator 2 (NS2) is presented. This module could be useful to speed-up the design, development and deployment of DiffServ-aware MPLS networks

This paper presents a simulative analysis of a modification to the TCP congestion control mechanism called ESSE (early slow start exit), designed to improve the TCP startup phase by setting the slow start threshold according to a pipesize estimation based on the observation of few ACK arrival times. We evaluate the performance of ESSE by using various methods to estimate the pipesize as the ratio between the round trip time and the spacing between ACK. This algorithm is easy to implement and preserves the compatibility with the standard protocol since it requires changes to the sender side only. Simulative experiments show that ESSE allows us to speed-up TCP connections and drastically reduce the packet drop rate under several working conditions and load levels. Better performance of TCP can be observed for any of the considered estimators, which indicates that the algorithm is robust against estimation errors. Further, the characteristics of fairness and friendliness (towards Newreno) of the algorithm are investigated. According to our simulations, ESSE-modified protocols guarantee fair utilization of bandwidth among homogeneous and heterogeneous (Newreno) connections sharing a common link.

The paper proposes an architecture for a scheduling algorithm to be integrated in IEEE 802.11 access points (AP), able to take into account, besides the transport service class required by the destination user, the channel quality experienced by the destination mobile station (STA). The relevance of this topic is due to the observation that when one or more STAs experience poor radio channel conditions, they increase the time spent to transmit a single packet, due to the retransmission of corrupted frames and the adoption of transmission techniques at a lower bit rate, leading to an inefficient use of the shared medium. These phenomena give, as a consequence, a worsening of the performance of all the STAs sharing the wireless medium, independent of their radio channel conditions. The adoption of a scheduling algorithm able to manage information on channel quality permits these effects to be reduced, not penalizing the STAs experiencing good channel conditions and, as a consequence, their experienced throughput. As a result, only the STAs in bad channel conditions experience a reduction of throughput. A prototype of an AP equipment implementing the proposed architecture is presented: it was obtained by modifying the hierarchical token bucket (HTB) and is referred to as wireless hierarchical token bucket (WHTB). The performance of the presented prototype is evaluated experimentally and compared with those obtained with standard scheduling algorithms, which do not take into account information on channel quality.

Learning requires a dialogue between abstract knowledge, such as theory, and application in specific contexts. A learning object (LO) for learning about network technology, scalable for level of complexity, and usable for abstractions in both business and engineering contexts, was developed through the application of two principles: separation of application from abstraction, and instructor-guided abstraction. The LO, design principles, and design and evaluation strategies, are described in this paper.

This paper presents the results of an experimental activity concerning the implementation and validation of an EF-PHB service in a high-speed metropolitan optical network with a differentiated services architecture. As EF-PHB can be used to create low loss, low latency and assured bandwidth services, real-time traffic flows have been aggregated and classified as EF. Measurements have been performed in different operating conditions to evaluate how the setting of the EF class of service parameters affects the time metrics of each traffic flow belonging to the aggregate. The transfer delay and the delay variation values, as specified by the ITU-T recommendation Y.1541 for real-time, jitter sensitive applications, have been taken as QoS performance objectives to satisfy.

The fractal nature of Internet traffic allows extending the application of the nonlinear chaotic system theory to the traffic control in modern telecommunication networks. In particular, the prediction techniques developed for these systems provide teletraffic engineers with a novel powerful tool for designing optimized traffic control algorithms. In this framework, the paper presents the performance evaluation of the Radial Basis Function Predictor (RBFP) in predicting actual traffic data. Predictor parameters are selected automatically by minimizing a suitably defined metric of prediction accuracy. The prediction system is then exploited in a simple resource allocation strategy to test the performance improvement achievable whenever a prediction of the future traffic intensity is available. The results obtained by means of discrete event simulation using actual traffic data are encouraging and stimulate a further-investigation of this approach.

In recent years, the development of a multiservice network, using integrated multiplexing and switching strategies, has converged to an IP-based infrastructure, reducing the relevance of other network solutions, such as a B-ISDN based on ATM. However, to develop an IP-based commercial network, the issues related to quality of service (QoS) support must be addressed. These problems are particularly relevant when a satellite access solution is considered, since the network resources can be limited and expensive. At the same time, demand assignment multiple access (DAMA) protocols permit better exploitation of satellite resources by dynamically allocating system bandwidth in response to user requests. The paper discusses the application to a DAMA satellite access network of different QoS architectures proposed by the IETF and the scientific community. The main novelty of the paper consists in selecting from the alternatives the most suitable solution to be applied to a satellite access network. Furthermore, the advantages and the drawbacks of the proposed architectures are analyzed in detail; the interaction between the control plane of the MAC layer and the resource management at the IP layer is highlighted.

The General Packet Radio Service (GPRS) networks are already implemented by several companies in different countries, however some problems related to the quality of service (QoS) observed by users and the efficient use of radio resources are not yet adequately addressed. In this framework, we propose a scheduling algorithm implemented at MAC layer, named as channel quality aware scheduler (CQAS). The main feature of the CQAS is that the user to serve is established taking into account the estimated channel quality for each of the active traffic flows. The performance of the proposed algorithm are evaluated in terms of goodput experimented by TCP clients, considering a set of concurrent file transfers and independent time varying channels. The comparisons between the CQAS performance and those obtained using a simple work conserving round robin (WCRR) algorithm, highlight the improvements introduced by the proposed algorithm. Moreover, the analysis emphasizes that the appropriate choice of both the channel coding scheme (CS) and the scheduling algorithm is needed in order to increase the radio resource utilization.

The paper deals with the applicability of fractional AutoRegressive Integrated Moving Average (f-ARIMA) processes to traffic modelling. This study is suggested by the ability of f-ARIMA processes in separately capturing both the long range dependence (LRD) and short range dependence (SRD) features of actual traffic. Indeed, the f-ARIMA models permit one to produce synthetic traces with a well-controlled LRD/SRD component ratio. The practical relevance of this feature is tested analytically, by means of discrete event simulations, the queueing performance in several working conditions for two different actual-traffic data sets, respectively related to a LAN-to-LAN interconnection and a videoconference service. In particular, the latter data set, characterised by a strong "low frequency" SRD component, emphasises the improvement on queueing performance evaluation introduced by a low-order f-ARIMA (p,d,0) model with respect to a pure LRD one. On the contrary, the inadequacy of the f-ARIMA model is highlighted when considering the LAN traffic characterised by a quite complex SRD component.

The paper presents a comparison of several traffic models in order to investigate their adequacy to approximate the behaviour of actual traffic in a single server queue under realistic working conditions. Hence, considering queueing systems with finite buffers, models parsimoniously fitting the long range dependence (LRD) property are compared to intrinsically short range dependent models and to flexible ones, capturing both LRD and short range dependence (SRD). The paper highlights an optimistic queueing behaviour evaluated by means of models capturing only LRD property, when realistic buffer sizes are considered and a strong SRD correlation component is present. The simulation results point out that the circulant modulated Poisson processes (CMPP) and fractional autoregressive integrated moving average (p,d,0) (f-ARIMA(p,d,0)) are suitable models to estimate the performance of the queueing system under heavy load conditions. On the contrary, when the system works under low-medium load conditions, the relevance of the marginal distribution to queueing performance and the inaccuracy of the CMPP model on fitting this statistic in the case of actual traffic data lead to a reduction of the accuracy of this model on system performance evaluation.

We analyze the accuracy of four methods for measuring effective bandwidths (EBs). We point out advantages and drawbacks of the four estimators. We find that for finite time realizations of a process the measured effective bandwidth differs considerably from its analytical counterpart. We also show that increasing the trace length has little impact on the accuracy of the measurements. We explain this behavior as a consequence of the intrinsic characteristics of the observed trace. We subsequently introduce the notion of "empirical effective bandwidth" (EEB) as a measure of performance tailored to the actual values. We derive properties of the EEB which capture its behavior in the parameter space and we contrast these properties with the ones obtained for analytical effective bandwidths. Finally, we comment on the use of EEBs in the context of connection admission control.

The paper presents joint measurements of IP and ATM traffic over an MPOA based campus network. The search for parsimonious realistic modeling of heterogeneous aggregated traffic has led to the generalization of on/off modeling schemes. Even a single on/off pattern with Pareto-geometric distributions of the sojourn times in each state determines a hyperbolic decay of the autocovariance function corresponding to long range dependence (LRIB). Hence, adequate on/off models can fit real traces corresponding to the multiplexing of several heterogeneous connections. In the paper, the on/off behavior of the traffic is evidenced at the ATM level. The relevant feature of the traces is that either the on and the off sojourn times could be well approximated by a light tailed distribution. The observed LRD behavior is due to the presence of correlation in the sequence of on (as well as off) periods lengths. In order to confirm this hypothesis, different synthetic traces were considered, verifying the LRD behavior by means of wavelet analysis. Implications of such state correlations on queueing behavior are then compared by means of discrete event simulations.

The paper presents the analysis of the applicability of /spl alpha/-stable processes in traffic modelling. This study is suggested by the goodness of /spl alpha/-stable processes in capturing not only the long range dependence (LRD) of actual traffic, but also the heavy tailness of its marginal distribution. The relevance of this property is proved by means of discrete event simulations carried out considering two different data traffic sets, respectively related to a LAN-to-LAN interconnection and entertainment video service. Errors in the estimation of Hurst parameter and in the evaluation of queueing behaviour are highlighted either analytically and empirically by simulations. In particular the queueing simulations have emphasised the improvements in the performance forecasting introduced by the higher flexibility of /spl alpha/-stable model with respect to the widely used fractional Brownian motion (FBM).

Preventive actions have to be used to avoid congestion in ATM networks. These actions are represented by call admission control (CAC) and usage parameter control functions (UPC). Two UPC algorithms are compared and an enhancement of the leaky bucket mechanism is presented. The comparison is based on the results of several trace driven simulations. The simulations, realised on a CAMAD tool, are driven by MPEG1 video traces as input traffic. The effectiveness of the proposed algorithm has been evaluated in terms of response time and transparency. These performance indexes have been obtained considering the number of non-conforming cells and the mean value of the conforming cells path.