INFOCOM 2020 论文列表

In an online social network, users exhibit personal information to enjoy social interaction. The social network provider (SNP) exploits users' information for revenue generation through targeted advertising. The SNP can present ads to proper users efficiently. Therefore, an advertiser is more willing to pay for targeted advertising. However, the over-exploitation of users' information would invade users' privacy, which would negatively impact users' social activeness. Motivated by this, we study the optimal privacy policy of the SNP with targeted advertising business. We characterize the privacy policy in terms of the fraction of users' information that the provider should exploit, and formulate the interactions among users, advertiser, and SNP as a three-stage Stackelberg game. By carefully leveraging supermodularity property, we reveal from the equilibrium analysis that higher information exploitation will discourage users from exhibiting information, lowering the overall amount of exploited information and harming advertising revenue. We further characterize the optimal privacy policy based on the connection between users' information levels and privacy policy. Numerical results reveal some useful insights that the optimal policy can well balance the users' trade-off between social benefit and privacy loss.

LoRa, more generically Low-Power Wide Area Network (LPWAN), is a promising platform to connect Internet of Things. It enables low-cost low-power communication at a few kbps over upto tens of kilometers with a 10-year battery lifetime. However, practical LPWAN deployments suffer from collisions, given the dense deployment of devices and wide coverage area. We propose CoLoRa, a protocol to decompose large numbers of concurrent transmissions from one collision in LoRa networks. At the heart of CoLoRa, we utilize packet time offset to disentangle collided packets. CoLoRa incorporates several novel techniques to address practical challenges. (1) We translate time offset, which is difficult to measure, to frequency features that can be reliably measured. (2) We propose a method to cancel inter-packet interference and extract accurate feature from low SNR LoRa signal. (3) We address frequency shift incurred by CFO and time offset for LoRa decoding. We implement CoLoRa on USRP N210 and evaluate its performance in both indoor and outdoor networks. CoLoRa is implemented in software at the base station and it can work for COTS LoRa nodes. The evaluation results show that CoLoRa improves the network throughput by 3.4× compared with Choir and by 14× compared with LoRaWAN.

Distributed synchronous stochastic gradient descent (SGD) algorithms are widely used in large-scale deep learning applications, while it is known that the communication bottleneck limits the scalability of the distributed system. Gradient sparsification is a promising technique to significantly reduce the communication traffic, while pipelining can further overlap the communications with computations. However, gradient sparsification introduces extra computation time, and pipelining requires many layer-wise communications which introduce significant communication startup overheads. Merging gradients from neighbor layers could reduce the startup overheads, but on the other hand it would increase the computation time of sparsification and the waiting time for the gradient computation. In this paper, we formulate the trade-off between communications and computations (including backward computation and gradient sparsification) as an optimization problem, and derive an optimal solution to the problem. We further develop the optimal merged gradient sparsification algorithm with SGD (OMGS-SGD) for distributed training of deep learning. We conduct extensive experiments to verify the convergence properties and scaling performance of OMGS-SGD. Experimental results show that OMGS-SGD achieves up to 31% end-to-end time efficiency improvement over the state-of-the-art sparsified SGD while preserving nearly consistent convergence performance with original SGD without sparsification on a 16-GPU cluster connected with 1Gbps Ethernet.

Recent work has demonstrated that cutting-edge advances in deep reinforcement learning (DRL) may be leveraged to empower wireless devices with the much-needed ability to "sense" current spectrum and network conditions and "react" in real time by either exploiting known optimal actions or exploring new actions. Yet, understanding whether real-time DRL can be at all applied in the resource-challenged embedded IoT domain, as well as designing IoT-tailored DRL systems and architectures, still remains mostly uncharted territory. This paper bridges the existing gap between the extensive theoretical research on wireless DRL and its system-level applications by presenting Deep Wireless Embedded Reinforcement Learning (DeepWiERL), a general-purpose, hybrid software/hardware DRL framework specifically tailored for embedded IoT wireless devices. DeepWiERL provides abstractions, circuits, software structures and drivers to support the training and real-time execution of state-of-the-art DRL algorithms on the device's hardware. Moreover, DeepWiERL includes a novel supervised DRL model selection and bootstrap (S-DMSB) technique that leverages transfer learning and high-level synthesis (HLS) circuit design to orchestrate a neural network architecture that satisfies hardware and application throughput constraints and speeds up the DRL algorithm convergence. Experimental evaluation on a fully-custom software-defined radio testbed (i) proves for the first time the feasibility of real-time DRL-based algorithms on a real-world wireless platform with multiple channel conditions; (ii) shows that DeepWiERL supports 16x data rate and consumes 14x less energy than a software-based implementation, and (iii) indicates that S-DMSB may improve the DRL convergence time by 6x and increase the obtained reward by 45% if prior channel knowledge is available.

Network tomography is an attractive methodology for inferring internal network states from accumulated path measurements between pairs of monitors. Motivated by previous results that identifying all link metrics can require a large number of monitors, we focus on calculating the performance bounds of a set of interesting links, i.e., bound-based network tomography. We develop an efficient solution to obtain the tightest upper bounds and lower bounds of all interesting links in an arbitrary network with a given set of end-to-end path measurements. Based on this solution, we further propose an algorithm to place new monitors over existing ones such that the bounds of interesting links can be maximally tightened. We theoretically prove the effectiveness of the proposed algorithms. We implement the algorithms and conduct extensive experiments based on real network topologies. Compared with state-of-the- art approaches, our algorithms can achieve 2.2~3.1 times more reduction on the bound interval lengths of all interesting links and reduce the number of placed monitors significantly in various network settings.

There have been increasing interests in exploring the sensing capabilities of RFID to enable numerous IoT applications, including object localization, trajectory tracking, and human behavior sensing. However, most existing methods rely on the signal measurement either in a low multipath environment, which is unlikely to exist in many practical situations, or with special devices, which increase the operating cost. This paper investigates the possibility of measuring `multi-path-free' signal information in multipath-prevalent environments simply using a commodity RFID reader. The proposed solution, Clean Physical Information Extraction (CPIX), is universal, accurate, and compatible to standard protocols and devices. CPIX improves RFID sensing quality with near zero cost - it requires no extra device. We implement CPIX and study two major RFID sensing applications: tag localization and human behavior sensing. CPIX reduces the localization error by 30% to 50% and achieves the MOST accurate localization by commodity readers compared to existing work. It also significantly improves the quality of human behaviour sensing.

We study low-overhead uplink multi-access algorithms for massive Internet-of-Things (IoT) that can exploit the MIMO performance gain. Although MIMO improves system capacity, it usually requires high overhead due to Channel State Information (CSI) feedback, which is unsuitable for IoT. Recently, a Pseudo-Random Beam-Forming (PRBF) scheme was proposed to exploit the MIMO performance gain for uplink IoT access with uniform channel and load, without collecting CSI at the BS. For non-uniform channel and load, new adaptive beamselection and random-access algorithms are needed to efficiently utilize the system capacity with low overhead. Most existing algorithms for a related multi-channel scheduling problem require each node to at least know some information of the queue length of all contending nodes. In contrast, we propose a new Low-overhead Multi-Channel Joint Channel-Assignment and Random-Access (L-MC-JCARA) algorithm that reduces the overhead to be independent of the number of interfering nodes. A key novelty is to let the BS estimate the total backlog in each contention group by only observing the random-access events, so that no queue-length feedback is needed from IoT devices. We prove that L-MC-JCARA can achieve at least `0.24`` of the capacity region of the optimal centralized scheduler for the corresponding multi-channel system.

In this paper, we study a novel Stochastic Network Utility Maximization (NUM) problem where the utilities of agents are unknown. The utility of each agent depends on the amount of resource it receives from a network operator/controller. The operator desires to do a resource allocation that maximizes the expected total utility of the network. We consider threshold type utility functions where each agent gets non-zero utility if the amount of resource it receives is higher than a certain threshold. Otherwise, its utility is zero (hard real-time). We pose this NUM setup with unknown utilities as a regret minimization problem. Our goal is to identify a policy that performs as `good' as an oracle policy that knows the utilities of agents. We model this problem setting as a bandit setting where feedback obtained in each round depends on the resource allocated to the agents. We propose algorithms for this novel setting using ideas from Multiple-Play Multi-Armed Bandits and Combinatorial Semi-Bandits. We show that the proposed algorithm is optimal when all agents have the same utility. We validate the performance guarantees of our proposed algorithms through numerical experiments.

The topology of a network is fundamental for building network infrastructure functionalities. In many scenarios, enterprise networks may have no desire to disclose their topology information. In this paper, we aim at preventing attacks that use adversarial, active end-to-end topology inference to obtain the topology information of a target network. To this end, we propose a Proactive Topology Obfuscation (ProTO) system that adopts a detect-then-obfuscate framework: (i) a lightweight probing behavior identification mechanism based on machine learning is designed to detect any probing behavior, and then (ii) a topology obfuscation design is developed to proactively delay all identified probe packets in a way such that the attacker will obtain a structurally accurate yet fake network topology based on the measurements of these delayed probe packets, therefore deceiving the attacker and decreasing its appetency for future inference. We show that ProTO is very effective against active topology inference with minimum performance disruption. Experimental results under different evaluation scenarios show that ProTO is able to (i) achieve a detection rate of 99.9% with a false alarm of 3%, (ii) effectively disrupt adversarial topology inference and lead to the topology inferred by the attacker close to a fake topology, and (iii) result in an overall network delay performance degradation of 1.3% - 2.0%.

Electric autonomous vehicles provide a promising solution to the traffic congestion and air pollution problems in future smart cities. Considering intensive energy consumption, charging becomes of paramount importance to sustain the operation of these systems. Motivated by the innovations in renewable energy harvesting, we leverage solar energy to power autonomous vehicles via charging stations and solar-harvesting rooftops, and design a framework that optimizes the operation of these systems from end to end. With a fixed budget, our framework first optimizes the locations of charging stations based on historical spatial-temporal solar energy distribution and usage patterns, achieving (2 + ε) factor to the optimal. Then a stochastic algorithm is proposed to update the locations online to adapt to any shift in the distribution. Based on the deployment, a strategy is developed to assign energy requests in order to minimize their traveling distance to stations while not depleting their energy storage. Equipped with extra harvesting capability, we also optimize route planning to achieve a reasonable balance between energy consumed and harvested en-route. Our extensive simulations demonstrate the algorithm can approach the optimal solution within 10-15% approximation error, and improve the operating range of vehicles by up to 2-3 times compared to other competitive strategies.

Traffic measurement is key to many network management tasks such as performance monitoring and cyber-security. Its aim is to inspect the packet stream passing through a network device, classify them into flows according to the header fields, and obtain statistics about the flows. For processing big streaming data in size-limited SRAM of line cards, many space-sublinear algorithms have been proposed, such as CountMin and CountSketch. However, most of them are designed for specific measurement tasks. Implementing multiple independent sketches places burden for online operations of a network device. It is highly desired to design a universal sketch that not only tracks individual large flows (called heavy hitters) but also reports overall traffic distribution statistics (called moments). The prior work UnivMon successfully tackled this ambitious quest. However, it incurs large and variable per-packet processing overhead, which may result in a significant throughput bottleneck in high-rate packet streaming, given that each packet requires 65 hashes and 64 memory accesses on average and many times of that in the worst case. To address this performance issue, we need to fundamentally redesign the solution architecture from hierarchical sampling to new progressive sampling and from CountSketch to new ActiveCM+, which ensure that per-packet overhead is a small constant (4 hash and 4 memory accesses) in the worst case, making it much more suitable for online operations, especially for pipeline implementation. The new design also makes effort to reduce memory footprint or equivalently improve measurement accuracy under the same memory. Our experiments show that our solution incurs just one sixteenth per-packet overhead of UnivMon, while improving measurement accuracy by three times under the same memory.

In multi-party interactive live streaming, each user can act as both the sender and the receiver of a live video stream. Designing adaptive bitrate (ABR) algorithm for such applications poses three challenges: (i) due to the interaction requirement among the users, the playback buffer has to be kept small to reduce the end-to-end delay; (ii) the algorithm needs to decide what is the bitrate to receive and what is the set of bitrates to send; (iii) the delay and quality requirements between each pair of users may differ, for instance, depending on whether the pair is interacting directly with each other. To address these challenges, we first develop a quality of experience (QoE) model for multi-party live streaming applications. Based on this model, we design MultiLive, an adaptive bitrate control algorithm for the multi-party scenario. MultiLive models the many-to-many ABR selection problem as a non-linear programming problem. Solving the non-linear programming equation yields the target bitrate for each pair of sender-receiver. To alleviate system errors during the modeling and measurement process, we update the target bitrate through the buffer feedback adjustment. To address the throughput limitation of the uplink, we cluster the ideal streams into a few groups, and aggregate these streams through scalable video coding for transmissions. We conduct extensive trace-driven simulations to evaluate the algorithm. The experimental results show that MultiLive outperforms the fixed bitrate algorithm, with 2-5× improvement in average QoE. Furthermore, the end-to-end delay is reduced to around 100 ms, much lower than the 400 ms threshold recommended for video conferencing.

Previous research on Segment Routing (SR) mostly focused on unicast, whereas online SDN multicast with segment trees supporting IETF dynamic group membership has not been explored. Compared with unicast SR, online SDN multicast with segment trees is more challenging since finding an appropriate size, shape, and location for each segment tree is crucial to deploy it in more multicast trees. In this paper, we explore Multi-tree Multicast Segment Routing (MMSR) to jointly minimize the bandwidth consumption and forwarding rule updates over time by leveraging segment trees. We prove MMSR is NP-hard and design an online competitive algorithm, named Segment Tree Routing and Update Scheduling (STRUS) to achieve the tightest bound. STRUS includes Segment Tree Merging and Segment Tree Pruning to merge smaller overlapping subtrees into segment trees, and then tailor them to serve more multicast trees. We design Stability Indicator and Reusage Indicator to carefully construct segment trees at the backbone of multicast trees and reroute multicast trees to span more segment trees. Simulation and implementation on real SDNs with YouTube traffic manifest that STRUS outperforms state-of-the-art algorithms regarding the total cost and TCAM usage. Moreover, STRUS is practical for SDN since its running time is about 1 second, even for massive networks with thousands of nodes.

Advances in mobile computing have paved the way for new types of distributed applications that can be executed solely by mobile devices on device-to-device (D2D) ecosystems (e.g., crowdsensing). More sophisticated applications, like cryptocurrencies, need distributed ledgers to function. Distributed ledgers, such as blockchains and directed acyclic graphs (DAGs), employ consensus protocols to add data in the form of blocks. However such protocols are designed for resourceful devices that are interconnected via the Internet. Moreover, existing distributed ledgers are not deployable to D2D ecosystems since their storage needs are continuously increasing. In this work, we introduce Mneme, a DAG-based distributed ledger that can be maintained solely by mobile devices and operates via two consensus protocols: Proof-of-Context (PoC) and Proof-of-Equivalence (PoE). PoC employs users' context to add data on Mneme. PoE is executed periodically to summarize data and produce equivalent blocks that require less storage. We analyze the security of Mneme and justify the ability of PoC and PoE to guarantee the characteristics of distributed ledgers: persistence and liveness. Furthermore, we analyze potential attacks from malicious users and prove that the probability of a successful attack is inversely proportional to the square of the number of mobile users who maintain Mneme.

In modern online social networks, each user is typically able to provide a value to indicate how trustworthy their direct friends are. Inferring such a value of social trust between any pair of nodes in online social networks is useful in a wide variety of applications, such as online marketing and recommendation systems. However, it is challenging to accurately and efficiently evaluate social trust between a pair of users in online social networks. Existing works either designed handcrafted rules that rely on specialized domain knowledge, or required a significant amount of computation resources, which affected their scalability.In recent years, graph convolutional neural networks (GCNs) have been shown to be powerful in learning on graph data. Their advantages provide great potential to trust evaluation as social trust can be represented as graph data. In this paper, we propose Guardian, a new end-to-end framework that learns latent factors in social trust with GCNs. Guardian is designed to incorporate social network structures and trust relationships to estimate social trust between any two users. Extensive experimental results demonstrated that Guardian can speedup trust evaluation by up to 2, 827 × with comparable accuracy, as compared to the stateof-the-art in the literature.

The rapid advancement of social media and communication technology enables video chat to become an important and convenient way of daily communication. However, such convenience also makes personal video clips easily obtained and exploited by malicious users who launch scam attacks. Existing studies only deal with the attacks that use fabricated facial masks, while the liveness detection that targets the playback attacks using a virtual camera is still elusive. In this work, we develop a novel video chat liveness detection system, LiveScreen, which can track the weak light changes reflected off the skin of a human face leveraging chromatic eigenspace differences. We design an inconspicuous challenge frame with minimal intervention to the video chat and develop a robust anomaly frame detector to verify the liveness of the remote user in the video chat using the response to the challenge frame. Furthermore, we propose resilient defense strategies to defeat both naive and intelligent playback attacks leveraging spatial and temporal verification. We implemented a prototype over both laptop and smartphone platforms and conducted extensive experiments in various realistic scenarios. We show that our system can achieve robust liveness detection with accuracy and false detection rates 97.7% (94.8%) and 1% (1.6%) on smartphones (laptops), respectively.

Different from using human-centric mobile devices like smartphones, unmanned aerial vehicles (UAVs) can be utilized to form a new UAV crowdsensing paradigm, where UAVs are equipped with build-in high-precision sensors, to provide data collection services especially for emergency situations like earthquakes or flooding. In this paper, we aim to propose a new deep learning based framework to tackle the problem that a group of UAVs energy-efficiently and cooperatively collect data from low-level sensors, while charging the battery from multiple randomly deployed charging stations. Specifically, we propose a new deep model called "j-PPO+ConvNTM" which contains a novel spatiotemporal module "Convolution Neural Turing Machine" (ConvNTM) to better model long-sequence spatiotemporal data, and a deep reinforcement learning (DRL) model called "j-PPO", where it has the capability to make continuous (i.e., route planing) and discrete (i.e., either to collect data or go for charging) action decisions simultaneously for all UAVs. Finally, we perform extensive simulation to show its illustrative movement trajectories, hyperparameter tuning, ablation study, and compare with four other baselines.

Network virtualization allows multiple tenant networks to coexist on a shared infrastructure. Core to its realization is the embedding of virtual networks onto the underlying substrate. Existing approaches are not suitable for cloud environments as they lack a fundamental requirement: elasticity. To address this issue we explore the capacity of flexibly changing the topology of a virtual network by proposing an embedding solution that adds elasticity to the tenant's virtual infrastructures. For this purpose, we introduce four primitives to tenants' virtual networks - including scale in and scale out - and propose new algorithms to materialize them. The main challenge is to enable these new services while maximizing resource efficiency and without impacting service quality. Instead of further improving existing online embedding algorithms - always limited by the inability to predict future demand - we follow a different approach. Specifically, we leverage network migration for our embedding procedures and to introduce a new reconfiguration primitive for the infrastructure provider. As migration introduces network churn, our solution uses this technique judiciously, to limit the impact to running services. Our solution improves on network efficiency over the state-of-the-art, while reducing the migration footprint by at least one order of magnitude.

Network function virtualization takes advantage of virtualization technology to achieve flexibility in network service provisioning. However, it comes at the cost of security risks caused by cache side-channel attacks on virtual machines. In this study, we investigate the security impact of these attacks on virtualized network functions. In particular, we propose a novel cache-based reconnaissance technique against virtualized Linux-based firewalls. The proposed technique has significant advantages in the perspective of attackers. First, it enhances evasiveness against intrusion detection owing to the ability of source spoofing. Second, it allows inference on a wide variety of filtering rules. During experiment in VyOS, the proposed method could infer the firewall rules with an accuracy of more than 90% by using only a few dozen packets. We also present countermeasures to mitigate cache-based attacks on virtualized network functions.