1,721,011 research outputs found
On the performance of TCP over a TDD-TD/CDMA architecture
No full textThis paper determines the performance of TCP when it is employed over a TDD-TD/CDMA architecture which supports different classes of subscribers and adopts an interference-driven admission control policy. The blocking and the outage probability of the system users are evaluated under various traffic conditions as well as the TCP throughput that data users experience. Call blocking and TCP throughput are determined by means of a semi-analytical approach. This work shows that when TCP is taken into account, many options come out for the system design in order to optimize radio resources. Resource assignment in the TDD-TD/CDMA system can then be done as a function of the actual radio link quality with the goal both to serve as many users as possible and to keep the TCP throughput as high as possible. Copyrigh
BBRp: Improving TCP BBR Performance over WLAN
Full text linkThis paper shows the inefficiency of TCP BBR in exploiting the Wi-Fi bandwidth. This limitation of BBR has been observed with both IEEE 802.11n and IEEE 802.11ac, where the mechanism of frame aggregation is used to boost the throughput of data transmission. In the last years, many TCP variants have been introduced to limit the bufferbloat phenomena and bound the latency through a reduction of the queue backlog injection rate. However, this mechanism impacts on the Wi-Fi frame aggregation logic, impeding TCP congestion controls to reach the full throughput potential of a Wi-Fi interface. While this problem can be solved with TCP Cubic by allowing the sender node to enqueue more packets, for TCP BBR the fix is not the same, as it has a customized pacing algorithm. With this contribution we propose BBRp, a new BBR version that allows for fine-tuning the congestion control pace, achieving between four and six times more throughput over IEEE 802.11n and IEEE 802.11ac channels, at the cost of an increased latency that is however always less than the latency obtainable with loss-based TCP congestion controls
The Impact of Transmission Power on the Safety-Related Performance of IEEE 802.11p
No full textIn this paper, we report and analyze results from a field test campaign aimed at assessing the real-world performance of IEEE 802.11p for safety-related applications. Regulator bodies in the EU and the US recommend the use of domestic-level transmission powers for general usage while granting the opportunity of employing higher powers for safety purposes. While the latter are generally expected to result in higher network performance, we want to quantitatively assess what they translate to in terms of metrics relevant from an application point of view, such as, goodput, round-trip times, jitter, and losses. These are studied and associated with conditions out of the network control at urban and suburban speeds, that is, in absence or in presence of different degrees of congestion, and in complete or partial line-of-sight. In general, we find evidence of an inverse correlation between the degree of congestion and the benefits granted by higher transmission powers. Up to the tested speed, IEEE 802.11p appears able to provide safety guarantees even at relatively long distances, as long as an appropriate transmission power is employed. At the same time, it must be acknowledged that higher power alone cannot overcome the significant dips in performance resulting from highly congested environments and non-line-of-sight scenarios
A hybrid algorithm to combine redundancy and concurrency in virtual network resource pooling
No full textEmergency scenarios pose significant hurdles to wireless communications, especially when these are performed through terrestrial cellular networks. This is because network infrastructures can be damaged, or the sudden surge of network demand can cause unbearable network congestion. Unfortunately, public safety operators employ terrestrial cellular networks to perform data communications and are therefore at risk of being unable to effectively operate in critical situations. Virtual network resource pooling is a software-defined networking (SDN)-based network framework that allows collaborative hosts to pool together their network channels to provide connectivity despite the sudden unavailability of network paths or to compensate for high degrees of packet losses. Once hosts’ resources are pooled, a dispatching algorithm on the SDN controller is required to enforce a proper policy of packets distribution. This paper presents a dispatching algorithm that is designed to provide fast and reliable transmissions despite lossy and unreliable channels. It works by introducing the minimum amount of packet redundancy that is needed to obtain a packet delivery probability given as a parameter and by using the remaining network resources to augment transfer goodput. Emulation results confirm that with respect to policies that selectively aim to improve goodput or to introduce packet redundancy, this hybrid algorithm can compensate for high packet loss ratios and improve goodput at the same time
Multi-Frequency Aeroelastic ROM for Transonic Compressors
Full text linkThe accurate prediction of the aeroelastic behavior of turbomachinery for aircraft propulsion poses a difficult yet fundamental challenge, since modern aircraft engines tend to adopt increasingly slender blades to achieve a higher aerodynamic efficiency, incurring an increased aeroelastic interaction as a drawback. In the present work, we present a reduced order model for flutter prediction in axial compressors. The model exploits the aerodynamic influence coefficients technique with the adoption of a broadband frequency signal to compute the aerodynamic damping for multiple reduced frequencies using a single training simulation. The normalized aerodynamic work is computed for a single oscillation mode at three different vibration frequencies, comparing the outputs of aerodynamic input/output models trained with a chirp signal to those from single-frequency harmonic simulations. The results demonstrate the ability of the adopted model to accurately and efficiently reproduce the aerodynamic damping at multiple frequencies and arbitrary nodal diameters with a single simulation
Impact of assembly algorithms on end-to-end performance in Optical Burst Switched networks with different QoS classes
No full textThis paper investigates the performance of a complete OBS (Optical Burst Switched) network that implements the JET (Just Enough Time) reservation mechanism. The network under examination adopts a core node architecture with no fiber delay lines and a limited set of wavelength converters, while featuring an edge node architecture with a mix of input, output and shared buffers. We investigate the overall network performance and design, specifically focusing on burst delay and end-to-end TCP throughput. In order to study the OBS network behavior, we develop a modular and flexible simulation tool, that we call MOBSSIM (Modular OBS Simulator). MOBSSIM is the means to accurately build an arbitrary topology OBS network via its main functional blocks, edge and core routers: its sound degree of development enables us to accurately evaluate several figures of merit, e.g., burst blocking probability and delay, and also allows for a critical comparison of alternative design solutions in terms of assembly algorithms, reservation strategies and QoS oriented routing
Performance evaluation of an interference based admission control scheme in T/CDMA systems with voice and data users
No full textThis paper analyzes the behavior of two time-code division architectures serving voice and data applications on the uplink of a cellular system. Both solutions attribute speech users one code over a time slot, whereas data users are granted several codes over the same time slot or a single code over several time slots. Call admission control is performed by a channel assignment algorithm which dynamically attributes resources if specified levels of transmission quality are met. The blocking probability of the two classes of users is determined, and it is demonstrated that the multicode T/CDMA approach performs better than the multislot solution and than a pure CDMA system
Comparing different GPS solutions for the Cooperative Awareness Messages Distribution
No full textCooperative awareness messages (CAM) are a family standardized by ETSI to distribute relevant information to all the nodes in vehicle coverage, such as time, position, speed, direction, etc. The parameters needed to trigger these messages, as well as the information contained in the messages themself, can be obtained with a GPS receiver. Our paper aims to increase the efficiency of CAM generation and dissemination in a less-is-more style, considering different GPS devices of different qualities. We achieve this goal by comparing the effectiveness of two GPS, belonging to two different market segments, with an order of magnitude of a price gap between them. We compared them in two different scenarios, similar in layout and traffic behavior but different in terms of GPS coverage conditions.Field trials with DSRC IEEE 802.11p have been carried out in an urban environment where different GPS coverage conditions widely emerge, with the possible presence of large buildings that emphasize all the GPS challenges: reducing the number of visible GPS satellites, introducing multi-path effects, fading, and reducing the signal strength.While having some notable differences in the two scenarios, we found the triggers' distributions to be similar between both the receiver devices. Instead, the average time between consecutive awareness messages has been found to be significantly lower in the case of bad coverage compared with the open-sky scenario, something that may reflect the presence of unwanted messages triggered by the GPS errors rather than the real vehicle motion. Therefore, we designed a new CAM generation algorithm that also considers the GPS error to reduce the messages transmitted on the radio medium
Quantifying Sleep-Related Energy Savings in Indoor LTE HetNets Radio Access
No full textWhile Sleep techniques show great promises in reducing the energy consumption of mobile networks, the necessity
of providing “always-on” services with current cellular architectures significantly hinders the introduction of effective Sleep
modes. As the process of network densification goes forward, it is paramount to lower to a minimum the additional energy
consumption required by the deployment of more and more network sites. Fortunately, the deployment of Heterogeneous
Networks opens new possibilities for the integration of simpler and highly effective Sleep technologies. In this paper, we evaluate
and quantify the energy savings attainable by the latest Power Modulation and Sleep modes in a Heterogeneous LTE Network
deployed to serve a traffic-intensive urban office area. Power consumption figures are based on those of real cells currently
available on the market. Results indicate that the combination of Power Modulation and Sleep techniques is able to cut in half
the energy required by the modeled Heterogeneous Network and that Sleep modes, in particular, can be the most beneficial
The Bufferboost Effect: when Drops and Redundancy boost the Throughput
No full textIn this paper, we present and characterize a phenomenon that arises when network congestion is met with packet
redundancy at a distributed bottleneck. Congestion typically causes packet drops at the queueing level, especially when Active
Queue Management techniques are employed to mitigate the bufferbloat effect. Redundancy may be present in the form of
packet replication over different paths, as a solution to increase network resilience and network availability guarantees. When
both network congestion and packet redundancy are in place, a counterintuitive throughput-boosting effect may originate from
them. We named this phenomenon bufferboost, for the key role played by packet drops introduced to avoid bufferbloat effects.
The contributions of this paper are the definition of the bufferboost phenomenon, its modelling through a mathematical upper
bound formulation, its isolation in an emulated environment for reproducibility and its validation through an extensive numerical
evaluation that also verifies the introduced model. Results show that bufferboost is a favorable side effect able to boost network
throughput
- …
