Search

Fog computing is an epitome to enable provisioning resources and services beyond the cloud, at the edge of the network, and nearer to end devices. Fog computing is not a counterfeit for cloud computing but a persuasive counterpart. However, the mobility of end users and/or fog nodes brings a major dilemma in the implementation of real-life scenarios. Therefore, we deliver the state-of-the-art research about mobility in fog computing. By identifying the mobility problems, requirements, and features of different proposals, we discover the open problems from subsisting studies and summarize the advantages of mobility for readers. This will help the researchers and developers avoid the current misapprehensions and capture the real-life scenarios, such as business, government, and education institutions. In spite of the extensive state-of-the-art research work, we also present the diverse mobility factors to investigate the correlation between dynamic nodes with fog nodes in order to accomplish better successful tasks. We conceive that the investigations of mobility factors in the fog computing will furnish the novel perspective on not only dynamic connections but also network dynamics. To revolutionize the follow-up research, we distinguish and foreground futurity directions concerning real-life scenarios of people and vehicles in a dynamic fog environment.

In recent years, much research effort on recommendation has been devoted to mining user behaviors, i.e., collaborative filtering, along with the general information which describes users or items, e.g., textual attributes, categorical demographics, product images, and so on. Price, an important factor in marketing --- which determines whether a user will make the final purchase decision on an item --- surprisingly, has received relatively little scrutiny. In this work, we aim at developing an effective method to predict user purchase intention with the focus on the price factor in recommender systems. The main difficulties are two-fold: 1) the preference and sensitivity of a user on item price are unknown, which are only implicitly reflected in the items that the user has purchased, and 2) how the item price affects a user's intention depends largely on the product category, that is, the perception and affordability of a user on item price could vary significantly across categories. Towards the first difficulty, we propose to model the transitive relationship between user-to-item and item-to-price, taking the inspiration from the recently developed Graph Convolution Networks (GCN). The key idea is to propagate the influence of price on users with items as the bridge, so as to make the learned user representations be price-aware. For the second difficulty, we further integrate item categories into the propagation progress and model the possible pairwise interactions for predicting user-item interactions. We conduct extensive experiments on two real-world datasets, demonstrating the effectiveness of our GCN-based method in learning the price-aware preference of users. Further analysis reveals that modeling the price awareness is particularly useful for predicting user preference on items of unexplored categories., 12 pages, 6 figures, to appear in ICDE2020

Cloud applications require the ability to customize bandwidth and network policies for desired wide area network (WAN) connections. Unfortunately, currently widely used virtual private network is difficult to achieve such customization where sophisticated manual configurations and operator expertise are required. Software-defined network (SDN) creates the opportunities to provide this ability. In this paper, we design Grace, a SDN-based system to provide diverse connections with flexible bandwidth and customized policies, implementing WAN as a service. Our contributions can be listed as follows: 1) We introduce open APIs for customers by abstracting WAN connections based on connection types, bandwidth, latency sensitivity, and policy-related information; 2) We develop an effective conflict detection algorithm considering both resource reservation and safety guarantee; 3) We propose a linear programming - based bandwidth algorithm for latency-sensitive connections by dynamic scheduling of time and bandwidth, and design a pricing scheme for various connection demands to address the case that network cannot fulfill all the connection requests. A prototype implementation and extensive evaluations show that Grace provides the ability to customize WAN connections without policy conflicts, allocates required bandwidth optimally, translates them into low-level configurations for underlying network devices, and successfully deploys WAN in a short time.

As the largest microblogging service in China, Weibo is not only an important social media but is also a source of human mobility data through its check-in service. We investigate the relationship between human mobility, Weibo accesses and check-in events using a large scale Weibo access dataset. Our findings indicate that the places where subscribers access Weibo only cover a small portion of their visited locations, and the number of Weibo accesses exhibits a long tail effect, indicating that some subscribers browse Weibo frequently. Check-ins not only deviate from individuals' daily routines, but they also provide a biased view of the actual location of subscribers. Therefore, relying on check-ins to infer human mobility patterns would lead to incorrect sociological conclusions.

Most existing recommender systems leverage user behavior data of one type, such as the purchase behavior data in E-commerce. We argue that other types of user behavior data also provide valuable signal, such as views, clicks, and so on. In this work, we contribute a new solution named NMTR (short for Neural Multi-Task Recommendation) for learning recommender systems from user multi-behavior data. In particular, our model accounts for the cascading relationship among different types of behaviors (e.g., a user must click on a product before purchasing it). We perform a joint optimization based on the multi-task learning framework, where the optimization on a behavior is treated as a task. Extensive experiments on the real-world dataset demonstrate that NMTR significantly outperforms state-of-the-art recommender systems that are designed to learn from both single-behavior data and multi-behavior data.

Fog computing is a promising solution for new emerging applications requiring intensive computation resources and low latency. Devices at the edge of network can share idle resources and collaboratively accomplish the computing tasks in fog computing. Thus, task publishers have heterogeneous options when offloading computing tasks considering the quality of transmission links, energy consumption and other hardware constraints of fog nodes. To incentivize these devices to participate in computation offloading, effective incentive mechanisms are needed. In this paper, utilizing the framework of contract theory, we formulate the negotiation between task ublisher and fog nodes as an optimization problem. The optimal contract is the Nash equilibrium solution achieved by task publisher and fog nodes. Simulation results show that an optimal contract can maximize the utility of task publisher meanwhile guarantee the individual rationality and incentive compatibility of fog nodes. Therefore, edge devices can be incentivized effectively to involve in the computation offloading.

Due to the proliferation of mobile networks and services, accessing online video services via mobile devices becomes increasingly popular, which generates huge data traffic and increases the cellular network load. It is valuable to characterize the usage patterns of different video services for network optimization and user experience enhancement. In this paper, we adopt a data-driven approach to investigate the usage patterns of three newly popular and major kinds of mobile video services: personalized livestreaming, user- generated video service, and traditional video portals. Based on the empirical analysis of a large dataset including 0.45 million users and 25 million logs, we find that 1) users have high loyalty in the same type of video services; 2) difference on the traffic peaks in consecutive days exists among different kinds of services; 3) personalized livestreaming services contributes a larger portion of video traffic, which still increases after midnight at weekday in the downtown. Finally, based on these findings, we discuss their implications and insights for network optimization and video service enhancement.

In the modern information society, analysis of human mobility becomes increasingly essential in various areas such as city planning and resource management. With users’ historical trajectories, the inherent patterns of their movements can be extracted and utilized to accurately predict the future movements. Plenty of previous work adopted traditional Markov model, which suffers when the trajectory becomes sparse or it shows distinct mobility patterns in different time of day. In this paper, based on an app-collected dataset of 100,000 individuals’ actively uploaded location information, we comprehensively analyze the mobility and predictability of each user. To approach the theoretical predictability and overcome the shortcomings of traditional Markov model, we propose a time-variant Markov model based on Gibbs sampling for mobility prediction. Specifically, we model human mobility as several interconnected Markov chains, each chain corresponds to a movement pattern of a period of time. Then, we adopt Gibbs sampling method to simultaneously recover the missing part of trajectories and train the Markov chains, in order to solve the unevenly distribution and the high missing rate. Results show that our prediction algorithm can achieve 11.2 percent higher prediction accuracy than the benchmark method, especially on sparse trajectories. In addition, we discover a high correlation between prediction accuracy and predictability, with correlation coefficient reaching 0.81. Finally, we investigate various factors including spatial and temporal resolution, orders of Markov models, and radius of gyration, in order to further explore the predictability under different circumstances.

Nowadays, human trajectories are widely collected and utilized for scientific research and business purpose. However, publishing trajectory data without proper handling might cause severe privacy leakage. A large body of works are dedicated to merging one's trajectory with others', so as to avoid any individual trajectory being re-identified. Yet their solutions do not provide enough protection since they cannot prevent semantic attack, which means the attackers are able to acquire individual's private information by using the semantics features of frequently- visited locations in the trajectory even without re- identification. In this work, we are the first to recognize the semantic attack, which is another severe privacy problem in publishing trajectory datasets. We propose an algorithm providing strong privacy protection against both the semantic and re- identification attack while reserving high data utility. Extensive evaluations based on two real-world datasets demonstrate that our solution improves the quality of privacy protection by 3 times, sacrificing only 36% and 30% of spatial and temporal resolution, respectively.

Enhancing throughput gain is an important target in the development of millimeter-wave (mmWave) communications. Spatial reuse is an important approach for throughput enhancement, which works well in mmWave networks due to directional links. In addition, multicast transmissions are highly desirable in mmWave systems, and become feasible with recent advances in antenna technology. In this paper, we address the problem of spatial reuse scheduling with multicast transmissions for mmWave communications. We formulate the problem as a Mixed Integer Non-Linear Programming (MINLP) problem and propose a low-computational-complexity heuristic algorithm, where potential links are fully utilized to exploit spatial reuse gain by dividing multicast demands into multiple unicast/multicast transmissions. Extensive simulation results demonstrate that the proposed scheme achieves near-optimal performance, and obtains high throughput gain with low packet loss rate.

For mobile networks diverse usage scenarios have different capability requirements on connection density and user experienced data rate, and modeling such capability diversity is crucial to the strategy evaluation in addressing the problem of high traffic load and scalability of network resources. Therefore, it is necessary to build a capability model in two dimensions of connection density and user experienced data rate. This paper aims at addressing this challenge based on an investigation of network capability in large-scale urban environment. First, our statistical study shows that the spatial distribution of these two parameters can be accurately fitted by log-normal mixture model. Second, we find that only six basic capability patterns exist among the 9,000 cellular base stations. Their connections with the urban functions of geographical locations are also explored in our work. Based on these two discoveries, we build a network capability model which can generate synthetic base stations with diverse connection density and user experienced data rate. We believe that this flexible and powerful model can help telecommunication operators to design and standardize mobile network in the future.

With the explosive growth of mobile data demand, small cells densely deployed underlying the homogeneous macro-cells are emerging as a promising candidate for the fifth generation (5G) mobile network. The backhaul communication for small cells poses a significant challenge, and with huge bandwidth available in the mmWave band, the wireless backhaul at mmWave frequencies can be a promising backhaul solution for small cells. In this paper, we propose the Maximum QoS-aware Independent Set (MQIS) based scheduling algorithm for the mmWave backhaul network of small cells to maximize the number of flows with their QoS requirements satisfied. In the algorithm, concurrent transmissions and the QoS aware priority are exploited to achieve more successfully scheduled flows and higher network throughput. Simulations in the 73 GHz band are conducted to demonstrate the superior performance of our algorithm in terms of the number of successfully scheduled flows and the system throughput compared with other existing schemes.

With the explosive growth of mobile data demands for local services, device-to-device (D2D) communication is proposed as a vital technology for next generation cellular network to enhance the transmission capacity. With the popularity of content-based mobile applications, popular content downloading can be optimized by allowing D2D communications to share the content with each other. Such improvements depend on effective and efficient solutions for the D2D technical problems of mode selection and resource allocation. Existing works dealing with these technical problems usually stand in the viewpoint of some network snapshots with fixed topology of static mobile users to optimization the instantaneous resource utilization. In this paper, from an angle of dynamic networks, we consider the mode selection and resource allocation problems for D2D communication underlaying cellular network with opportunistic change of users' location by fully exploiting their mobility patterns to achieve global optimization. By formulating the above process-oriented optimization problems, we introduce a utility function and contention graph based approaches, termed MAD2D (Mobility assisted D2D), to heuristically solve them effectively. Through extensive simulations, we demonstrate MAD2D greatly enhanced the system performance compared with the state-of-the-art solutions.

In the future super dense wireless networks, millimeter-wave (mmWave) communications systems have great prospects, mainly due to the huge bandwidths and the directional transmissions. However, phase noise is significant due to the high oscillation frequency, which affects channel estimation and deteriorates the bit-error-rate (BER) performance. This paper emphasizes that phase noise may be estimated in the frequency domain, and a scheme with iterative channel estimation and phase noise compensation is proposed for single-carrier frequency-domain equalization (SC-FDE) based mmWave systems. We achieve channel estimation by calculating small perturbations iteratively with the first-order approximation. For signal demodulation, we adopt an iterative receiver to compensate phase noise with the decision feedback result. Comprehensive simulations demonstrate that our scheme achieves competitive performance and outperforms the traditional methods, in terms of both mean square error (MSE) in channel estimation and BER in signal demodulation.

Due to the requirement of establishing several independent links, multi-device multi-path beamforming (BF) training becomes rather time-consuming in 60-GHz millimeterwave communications. To reduce training overhead, a novel BF training scheme is proposed in this paper. The scheme includes a simultaneous training strategy for multiple devices and multiple paths, and a corresponding multi-path beam-pair detection algorithm. The training pilot of the source device is broadcasted to all destination devices. Subsequently, in each destination device, the proposed detection is separately conducted to detect significant-path-related beam pairs from the channel responses, via exploiting the time resolvability and space sparsity of physical paths. Finally, the training results are successively fed back to the source device. The proposed scheme achieves simultaneous training of multiple devices and multiple paths, and obtains significant reduction of overhead. Comprehensive simulations are presented to verify the effectiveness of this scheme.

The performance of wide area network (WAN) has a direct impact on user experience of cloud usage. In order to provide on-demand and performance-assured WAN connections, we design and implement a SDN-based system called Grace, which abstracts the underlying network resources into life-cycle network services to provide the predefined network performance via open APIs. With the advantages of this system, we demonstrate that WAN as a service for cloud can be achieved in event of the deployed Grace.

With tremendous growth in demand for broadband data, we undergo severe traffic congestion problems in mobile data networks. Researchers found out that there exists spatio-temporal inhomogeneity in data traffic, which lacks a measurement-driven study. In this work, by carrying out a thorough analysis of data traffic obtained from 380,000 base stations in a large-scale metropolis of Shanghai, we quantitatively characterize the “tide” phenomenon in traffic usage. In different days, its distribution shows an exponential decay, while in different places it follows a power law distribution. These findings would provide important insights on how to deal with traffic congestion problems.

Various network optimization and management goals in cloud data centers can be achieved by re-mapping VMs to the substrate servers. Live VM migration is used to implement such re-mapping by moving VMs from the initial servers to the target ones. For efficiency and usability considerations, these migration tasks are expected to be completed as soon as possible, which can be achieved by planning multiple VMs to be migrated simultaneously. However, the available resources of computation and bandwidth limits the number of VMs that can be migrated at the same time. Besides, the available resources change with the progress of migration process, which makes VM migration planning a challenge problem. Considering both resources constraints and the dependence between the migration plan and the total migration time, in this paper, we investigate the problem of virtual machine migration planning to minimize the total migration time. We formulate VM migration planning as an optimization problem considering both computation and bandwidth constraints. Our formulation is based on a step-by-step migration scheme, where multiple VMs can be migrated simultaneously as long as the resource constraints are satisfied. The solution of the optimization problem outputs the migration plan which achieves the minimum time to finish the migration task. Moreover, we verify the effectiveness of the optimal migration plan through extensive simulations and comparison with the best algorithm in existing works.

With the significant growth of mobile data traffic in the coming years, 60-GHz millimeter-wave (mm-wave) communication shows great potential in high throughput networks to meet this demand. The directional transmission in 60-GHz radio has motivated studies of spatial sharing (SPSH). In IEEE 802.11ad, a widely acknowledged protocol of 60-GHz system, spatial sharing inside a network is supported. In addition, different networks are recommended to transmit in non-overlapping period to mitigate inter-network interference, which is hard to employ in real system under circumstances such as heavy traffic load. However, there is no specific mechanism for spatial sharing with interference mitigation among networks in IEEE 802.11ad standard. Therefore, we propose an inter-network spatial sharing strategy in IEEE 802.11ad Wireless Local Area Networks (WLANs) in this paper. Specifically, the strategy includes SPSH report, which helps establish a complete interference and coexistence database. Corresponding time scheduling scheme is based on mutual avoidance of interference. Extensive simulations have shown that our proposed strategy enhances spatial gain and depicted more robust performance compared with existing mechanism in IEEE 802.11ad.

In mobile delay tolerant networks (DTNs), energy-efficient probing schemes are vital to prolong the limited battery life of mobile devices. Current solutions mainly focus on optimizing the probing interval during contacts by assuming the contact rate between devices are exponentially distributed. However, extensive experimental measurements have shown the contact rate follows a power law distribution, which indicates the current solutions not optimal while on the other hand dramatically increases the difficulty to analyze this problem. In this paper, we propose an adaptive wakeup schedule based on power-law distributed contacts. Our approach requires nodes stay asleep when a contact is unlikely to happen and wake up only when the possibility that it can successfully contact another node is relatively high. By predicting node contacts in the future, our approach significantly reduce energy consumption without degrading the performance of opportunistic networks. Extensive simulations with real-life human traces have demonstrated the outstanding performance of our approach, which are able to save 30 percent energy while keeping the same performance in most scenarios.

Due to human blockage and device mobility, 60GHz wireless local-area network (WLAN) faces the challenge of communication-link outage. To cope with this problem, a double-link beam tracking scheme is proposed in this paper. This scheme includes a search strategy of transmission and alternative links, and an in-packet double-link tracking and switching method. The transmission and alternative links are simultaneously tracked against device mobility, through probing neighboring beams; once the human blockage is detected in transmission link, the system can thus switch to the alternative link directly. Besides, the proposed scheme finds the finish of human blockage quickly through in-packet double-link tracking, and enables the system to communication in the maximal received-power link. The comparison between the proposed scheme and the IEEE 802.11ad scheme is conducted. The comparison shows that the proposed scheme significantly reduces the probability of outage, and achieves an improved throughput.

Multicast services are highly desirable in 60-GHz millimeter-wave communications, but multicast beamforming remains a challenge due to special structure of high frequency system. To solve this problem, a transmit and receive multicast beamforming scheme has been proposed in this paper. Unlike conventional multicast models in LTE/EMBMS, every device in proposed system is equipped with a multi-antenna array, which helps enhance beam directivity and reduce interference. Two stages are put forward to generate transmit and receive beams for multicasting. Performance evaluations have shown that the total throughput of the network has been improved largely by using the proposed scheme compared with unicast beamforming. Additionally, the designed algorithm has a minor performance loss compared with previous quasi-optimal algorithm, but with much lower complexity, which makes it advantageous for practical use.

In millimeter-wave communications (MMWC), multipath components (MPCs) have different steering angles and independent fadings. Park and Pan recently proposed a simple scheme for both transmitter and receiver to concurrently beamform towards multiple steering angles of MPCs to achieve both array gain and diversity gain. However, when the number of MPCs is greater than that of the transmit or receive antennas, a solution of antenna weight vector (AWV) does not exist. To cope with this problem, two multipath grouping (MPG) schemes, namely MPG-steering vector grouping and MPG-channel vector grouping, are proposed. These schemes group the MPCs that have close steering angles, and define an equivalent MPC for each non-empty group in both the transmitter and receiver. As the number of non-empty groups is always no larger than that of antennas in both the transmitter and receiver, a solution of AWV is guaranteed. Moreover, performance comparisons show that the two proposed MPG schemes achieve not only full diversity, but also an even better array gain than that of the scheme proposed by Park and Pan.

Understanding and characterizing the properties of vehicular networks has become hugely important because of their wide applications and fast development. Due to the unique property of vehicular networks that nodes change their locations at high speed in anytime, their connectivity is an important and unique property. In this poster, based on realistic vehicular mobility traces, we reveal the opportunistic connectivity property of large-scale urban vehicular networks. Moreover, through analysis, we unveil the fundamental relationships and tradeoffs between the opportunistic connectivity and network parameters in terms of data size, delivery delay, transmission energy, etc.

Despite of various benefits such as load balance and energy saving virtual machine (VM) migration promises to provide, its application in realistic data centers is still limited due to the limitation of migration in the LAN environment and the unpredictable performance. Through separation of control plane and data plane, software defined network (SDN) provides the possibility for an alternate solution to overcome these limitations. In this work, we aim at designing and implementing a software defined approach for live VM migration to experiment the possibility. This paper presents the key challenges and our preliminary design result to address these challenges.

In this paper, a novel lower-power multigigabit Viterbi decoder architecture is proposed for 60 GHz wireless personal-area network (WPAN) systems. Since add-compare-select (ACS) computation is the main bottleneck in decoding speed of Viterbi decoders, a balanced add-select-register-compare (BASIC) architecture is put forward to reduce critical path of ACS with low complexity. This work develops an 8-parallel BASIC Viterbi decoder for IEEE 802.15.3c standard to accomplish high-throughput and low-power goals. Based on synthesized results in 0.13 μm CMOS technology, the proposed decoder achieves up to 4 Gb/s throughput with energy efficiency 0.104 nJ/bit at 1.2 V.

Knowledge of where vehicles will be in near future helps users in daily planning, traffic monitors in vehicles scheduling, advertisers in fixed point advertising, and especially helps in communication network source provisioning. In this paper, we analyze the predictability of taxi mobility based on their locations and time period records and we present a prediction method of taxis for their next locations in 15 seconds using Markov predictor. The historical location trace of each taxi is used to train the transition probability matrix of next location for our predictor, and we use 3 different scenarios to predict. Based on records from over 2,000 taxis in Shanghai, and over 14,000 taxis in Beijing, we are able to predict the next vehicular location with an accuracy of 82%.

Wireless virtualization has been proposed as a promising way to overcome the current plight of the wireless network, i.e. spectrum crisis, by enabling multiple concurrent wireless virtual networks running on shared wireless substrate resource. However, making efficient use of the underlying spectrum requires effectively assigning the physical spectrum resources to several virtual networks. Since opportunistic spectrum sharing makes wireless communication more intelligent and efficient, we present an opportunistic spectrum sharing based resource allocation scheme for wireless virtualization. First, we formulate the resource allocation problem as an NP-Hard integer program through employing opportunistic spectrum sharing. Then, we propose a dynamic programming algorithm and a heuristic algorithm to solve it. Simulation results confirm the advantage of our algorithms and show that opportunistic spectrum sharing significantly improves the revenue, resource utilization and accept rate while decreasing the payments of virtual networks.

Due to the extremely high oscillation frequency of 60 GHz systems, phase noise (PN) imported at both transmitter (Tx) and receiver (Rx) is significant, which degrades the transmission performance. This paper proposes an architecture employing iterative Tx and Rx PN compensation (ITR-PNC) for 60 GHz systems with single-carrier frequency-domain equalization (SC-FDE) transmission. The ITR-PNC iteratively performs PNC before equalization (PNC-BE) and PNC after equalization (PNC-AE), which are mainly set to manage Rx PN (RPN) and Tx PN (TPN), respectively. The PNC-BE and PNC-AE both exploit the one-tap least mean square (LMS) algorithm for PN extraction (PNE). In PNC-AE, the decision feedback result is used as the reference signal for PNE. And in PNC-BE, the reference signal is the signal replica, which is generated via the decision result and the estimated TPN of last iteration, and the estimated channel response. Comprehensive simulations indicate that the proposed architecture employing the proposed ITR-PNC achieves competitive bit-error-rate (BER) performance with only two iterations, for the cases whether only RPN or both TPN and RPN are taken into consideration.

Data plane programmability is regarded as an essential feature of future Internet. However, the flexibility of programmability has not been carefully considered. We present GrainFlow, a hardware-based packet processing platform supporting customization on every header bit's processing and adapting to packet header length variation. We evaluate GrainFlow on a Virtex-5 FPGA board. Experiment results show that users can flexibly redefine the data plane with only 10-20 rules and data plane forwarding rate can achieve wire-speed with acceptable latency increase.

Roadside Units (RSUs), which enable the vehicles-to-infrastructure communication, are deployed along the roadside to handle the growing communication demands as the number of vehicles increases. How to optimally deploy the RSUs and efficiently use them to enhance the vehicular network performance are important problems in the RSU-aided opportunistic networks. In order to solving these problems, we need to understand fundamental laws and properties of the opportunistic contacts. Due to the prohibitive costs to collect enough realistic contact records, to the best of our knowledge, little experiment work has been conducted to study the communication contact patterns in RSU-aided opportunistic networks. In this work, we carry out an extensive experiment involving tens of thousands of operational taxis in Beijing city. Based on this newly collected Beijing trace and the existing Shanghai trace, we find some invariant characters for the communication contacts for large scale RSU-aided vehicular opportunistic networks. In terms of inter-contact time between the RSUs and vehicles, we find it obeys exponential distribution, while in terms of contact rate, it follows Poisson distribution. Our observations reveal fundamental patterns for large scale vehicular mobility, and further provide useful guidelines for the design of new routing protocols and their performance evaluation for RSU-aided vehicular opportunistic networks.

Opportunistic contact between moving vehicles is one of the key features in vehicular delay tolerant networks (VDTNs) that critically influences the design of routing schemes and the network throughput. Due to prohibitive costs to collect enough realistic contact recodes, to the best of our knowledge, little experiment work has been conducted to study the opportunistic contact patterns in large scale urban vehicular mobility environment. In this work, we carry out an extensive experiment involving tens of thousands of operational taxis in Beijing city. Based on studying this newly collected Beijing trace and the existing Shanghai trace, we find some invariant characteristics of the opportunistic contacts for large scale urban VDTN. Specifically, in terms of contact duration, we find that there exists a characteristic time point, up to which and including at least 80% of the distribution, the contact duration obeys an exponential distribution, while beyond which it decays as a power law one. This property is in sharp contrast to the recent empirical data studies based on human mobility, where the contact duration exhibits a power law distribution. In terms of contact interval, we find that its distribution can be modelled by a three-segmented distribution, and there exists a characteristic time point, up to which the contact interval obeys a power law distribution, while beyond which it decays as an exponential one. Our observations thus reveal fundamental patterns for large scale vehicular mobility, and further provide useful guidelines for the design of new urban VDTN' routing protocols and their performance evaluation.

The direct-conversion transceiver is an attractive architecture for 60 GHz wireless systems to meet the low-cost and low-power-consumption requirements, but it suffers from the IQ imbalance problem seriously. This paper adopts secondorder moment estimation to extract frequency-independent IQ imbalance parameters, and performs IQ compensation based on the estimation results. In order to achieve robust estimation, we employ a specific preamble sequence, where adjacent symbols have a fixed phase difference of π/2. The proposed estimation excels in the following aspects: (1) it completely eliminates the impact of inter-symbol interference (ISI) and carrier phase offset (CPO); (2) it is not degraded by carrier frequency offset (CFO), since the impact of which only lies in the negligible phase offset during only a symbol duration; (3) it has low computational complexity and can be implemented on hardware platform expediently. Simulation results indicate that the estimation algorithm has competitive performance over a large range of IQ imbalance, and with IQ compensation the bit-error-rate (BER) performance can be improved to that of the ideal case.

Beamforming is necessary in 60GHz millimeterwave communications in order to compensate for high path loss. A complete beamforming algorithm includes codebook design and beam training. The existing beamforming protocol of IEEE 802.15.3c standard has following flaws: (1) Not all three beamforming pattern types (quasi-omni pattern, sector and beam) have their own codebooks; (2) Almost every beam points to two directions, one of which is expected but the other one is unwanted; (3) Setup time of beam training is long. This paper proposes a codebook design and corresponding training procedure based on circular antenna array with a two-layer-structure. It allows every pattern type to own its corresponding codebook and highly directional antenna radiation pattern. The training setup time is reduced to around half of 3c training time. Performance evaluations also manifest that the proposed algorithm leads to less antenna gain loss and more robust performance than the 3c standard.

The newly emerged vehicular communication network is seen as a key technology for solving the increasingly serious vehicular traffic congestion as well as improving road safety, and applications of vehicular networks are also emerging at the same time. Predicting vehicular visiting time is vital to both solving the vehicular system problem and building efficient vehicular networking. It is an open and unsolved problem that how much the vehicular staying duration of visits to different areas can be predicted. In this paper, we use real vehicular traces in Beijing and Shanghai and an area partition model to explore the limit of predictability of the vehicular visiting time in different areas in large scale cities and analyze the effect of different precision and slot time on predictability. We conclude that using a proper time slot is an efficient way of prediction and a higher predictability can be achieved if the requirement for precision is reduced. Among all the cases we study, we find the highest potential predictability in Beijing and Shanghai in the case with smallest slot time and lowest requirement for precision.

Modeling complicated vehicular mobility behavior and the associated network connectivity in large scale vehicular ad hoc network has been a challenging problem for the past several years. In this paper, we carry out extensive experiments involving thousands of operational taxies in Shanghai and Beijing cities. Studying the taxi traces on the vehicular connectivity, we observe that the vehicle isolation probability exhibits an exponential decay with the increase of transmission range. This observation is in sharp contrast to the most widely used system models such as Random Way Point and Random Walk, where the vehicle isolation probability changes with transmission range as exp(-λR2), where R is the transmission range. We propose a simple theoretical model to explain the exponential decay of vehicle isolation. Validation of our model is carried through comparisons between the synthetic trace and the real trace, and the results show that our mobility model has a good approximation to the real scenario. Our results give insights on the behavior of connectivity, thus providing fundamental guidelines on design of new vehicular mobility models, new data forwarding protocols and their performance analysis.

60 GHz communication supporting multigigabit data rate is a popular choice of industry for next generation short distance wireless communications. However, multi-Gsps ADC becomes a challenge in 60 GHz systems which have ultra wide Nyquist bandwidth. To reduce sampling rate of ADC in the estimation stage, we propose a joint signal-to-noise ratio (SNR) and channel estimation algorithm using compressed sensing (CS) theory. In 60 GHz systems, CS encoding and decoding strategies are optimized to maximize benefits from the design of pilots and estimators. For pilot design, m-sequence rather than conventional Bernoulli random sequence is selected owning to a better average restricted isometry property; for estimator design, a quasi-optimal channel and noise power estimation is put forward underlying signal subspace provided by CS algorithm. Simulation results show that the proposed algorithm reduces the sampling rate of ADC to 9.1% Nyquist bandwidth of 60 GHz communication. Moreover, the algorithm with this compressed sampling efficiently outperforms classical least square algorithm with Nyquist sampling as SNR exceeds 7 dB.

Currently, Delay Tolerant Networking (DTN) research community urgently needs a credible and flexible platform to setup large scale DTN networks to accurately evaluate the performance of algorithms and protocols. In this work, we present such an emulation testbed. By using network virtualization and OpenFlow technologies to generate virtual DTN nodes and to simulate the opportunistic links controlled by mobility scenarios, our designed testbed is a realistic DTN protocol evaluation environment composing realistic networks and applications, which provide high credibility, flexible controlling, and supporting for large scale experiments.

Vehicular Delay Tolerant Networks (VDTN) are used to distribute a large amount of mobile data by high-capacity device-to-device communication. Given a number of available vehicles in VDTN, the current vehicular data sharing models always utilize all of them to achieve the best effect, failing to balance the performance gained and cost of employing them. Taking the cost into account, we study the problem of the optimal number of vehicles to be deployed in the data sharing system, which considers using erasure codes to further increase the system efficiency if possible. By establishing the system goal which consists both the effectiveness and the cost of the vehicular network mathematically, we formulate the above problem as a utility function minimization problem. Finally, we solve the problem by theoretical derivation, and demonstrate the efficiency of the obtained solution through simulations using real vehicular traces of Beijing and Shanghai.

Wireless traffic has two features: people have common interests in the hottest contents, and a large amount of traffic belongs to non-realtime services. Exploiting these two features, we propose a content dissemination system, integrating broadcast network and opportunistic network. In the proposed system, popular contents are broadcasted through broadcast network and then disseminated to users through opportunistic network. Such system can significantly reduces the same contents repeated transmission and efficiently delivers contents to their clients. Using a Markov model, we theoretically analyze content dissemination delay of the proposed system.

Network virtualization provides an alternative for wireless testbed to support multiple experiments simultaneously, as its properties of isolation, customization and efficiency are able to improve network resources utilization significantly. However, complex radio environment makes the employment of virtualization in wireless networks become a difficult problem, in which strict isolation across different experiments is hard to maintain. What is more, present radio experimental systems mostly exploit only Time Division Multiplexing, which may cause resources waste when the scale of experiments is relatively small. In this paper, we propose a time-space combined resources allocation scheme by taking advantage of both Time Division Multiplexing and Space Division Multiplexing. In this scheme, we first select a new time slot and some unscheduled experiments. Then we adopt Space Division Multiplexing to fit in as many experiments as possible in a time slot while ensuring required isolation across them. Extensive simulation shows that our scheme improves experimental resources utilization effectively, while each experiment in the same time slot have enough bandwidth among all nodes and the fairness of resources utility is also ensured.