INFOCOM 2022 论文列表

The discovery of wireless power transfer technology enables power transferred between transceivers in a wireless manner, thus generating the concept of wireless rechargeable sensor networks (WRSNs). Previous arts paid little attention to network security issues, making them prone to novel attacks. In this work, we focus on developing a denial of charge attack for WRSNs, which aims at corrupting network functionalities by manipulating the malicious mobile charger. We formalize the maximization of destructiveness problem (MAD) and propose a denial of charge attacking method, termed MDoC, with a performance guarantee to solve it. MDoC is composed of two attacking rounds, which first triggers sensors to send requests to create a request explosion phenomenon and then figures out the longest charging route to yield nodes starving to death as much as possible. Finally, extensive testbed experiments and simulations are conducted to verify the performance of MDoC. The results reveal that MDoC attack is able to exhaust at least 20% additional nodes without being noticed.

Big client data and deep learning bring a new level of accuracy to wireless traffic prediction in non-adversarial environments. However, in a malicious client environment, the training-stage vulnerability of deep learning (DL) based wireless traffic prediction remains under-explored. In this paper, we conduct the first systematic study on training-stage poisoning attacks against DL-based wireless traffic prediction in both centralized and distributed training scenarios. In contrast to previous poisoning attacks on computer vision, we consider a more practical threat model, specific to wireless traffic prediction, to design these poisoning attacks. In particular, we assume that potential malicious clients do not collude or have any additional knowledge about the other clients’ data. We propose a perturbation masking strategy and a tuning-and-scaling method to fit data and model poisoning attacks into the practical threat model. We also explore potential defenses against these poisoning attacks and propose two defense methods. Through extensive evaluations, we show the mean square error (MSE) can be increased by over 50% to 108times with our proposed poisoning attacks. We also demonstrate the effectiveness of our data sanitization approach and anomaly detection method against our poisoning attacks in centralized and distributed scenarios.

While Deep Neural Network (DNN) models have transformed machine vision capabilities, their extremely high computational complexity and model sizes present a formidable deployment roadblock for AIoT applications. We show that the complexity-vs-accuracy-vs-communication tradeoffs for such DNN models can be significantly addressed via a novel, lightweight form of "collaborative machine intelligence" that requires only runtime changes to the inference process. In our proposed approach, called ComAI, the DNN pipelines of different vision sensors share intermediate processing state with one another, effectively providing hints about objects located within their mutually-overlapping Field-of-Views (FoVs). CoMAI uses two novel techniques: (a) a secondary shallow ML model that uses features from early layers of a peer DNN to predict object confidence values in the image, and (b) a pipelined sharing of such confidence values, by collaborators, that is then used to bias a reference DNN’s outputs. We demonstrate that CoMAI (a) can boost accuracy (recall) of DNN inference by 20-50%, (b) works across heterogeneous DNN models and deployments, and (c) incurs negligible processing, bandwidth and processing overheads compared to non-collaborative baselines.

Motivated by the popularity of edge intelligence, DNN services have been widely deployed at the edge, posing significant performance pressure on edge servers. How to improve the QoS of edge DNN services becomes a crucial and challenging problem. Previous works, however, did not fully consider the heterogeneous QoS requirements on urgent and non-urgent tasks, causing frequent QoS violations. Meanwhile, our empirical study shows that severe task interference exists in concurrent DNN tasks, further degrading the timeliness of urgent tasks and throughput of non-urgent tasks. To address these issues, we propose Kalmia, a heterogeneous QoS-aware framework for DNN inference task scheduling on edge servers. Specifically, Kalmia includes an offline profiling stage and an online scheduling policy. In offline profiling, we build a regression model to predict the execution time of tasks. During online scheduling, we classify the tasks into urgent and non-urgent tasks and distribute them into two CUDA contexts. By a tailored scheduling strategy, non-urgent tasks can fully utilize the computing resources for throughput improvement, while the timeliness of urgent tasks can be guaranteed via preemption. Experimental results demonstrate that Kalmia can achieve up to 2.8× improvement in throughput and significantly reduce the deadline violation rate compared with state-of-the-art methods.

Bulk synchronous parallel (BSP) is the de-facto paradigm for distributed DNN training in today’s production clusters. However, due to the global synchronization nature, its performance can be significantly influenced by network bottlenecks caused by either static topology heterogeneity or dynamic bandwidth contentions. Existing solutions, either system-level optimizations strengthening BSP (e.g., Ring or Hierarchical All-reduce) or algorithmic optimizations replacing BSP (e.g., ASP or SSP, which relax the global barriers), do not completely solve the problem, as they may still suffer from communication inefficiency or risk convergence inaccuracy.In this paper, we present a novel divide-and-shuffle synchronization (DS-Sync) to realize communication efficiency without sacrificing convergence accuracy for distributed DNN training. At its heart, by taking into account the network bottlenecks, DS-Sync improves communication efficiency by dividing workers into non-overlap groups to synchronize independently in a bottleneck-free manner. Meanwhile, it maintains convergence accuracy by iteratively shuffling workers among different groups to ensure a global consensus. We theoretically prove that DS-Sync converges properly in non-convex and smooth conditions like DNN. We further implement DS-Sync and integrate it with PyTorch, and our testbed experiments show that DS-Sync can achieve up to 94% improvements on the end-to-end training time with existing solutions while maintaining the same accuracy.

Southbound message delivery from the control plane to the data plane is one of the essential issues in multi-tenant clouds. A natural method of southbound message delivery is that the control plane directly communicates with compute nodes in the data plane. However, due to the large number of compute nodes, this method may result in massive control overhead. The Message Queue (MQ) model can solve this challenge by aggregating and distributing messages to queues. Existing MQ-based solutions often perform message aggregation based on the physical network topology, which do not align with the fundamental requirements of southbound message delivery, leading to high message redundancy on compute nodes. To address this issue, we design and implement VITA, the first-of-its-kind work on virtual network topology-aware southbound message delivery. However, it is intractable to optimally deliver southbound messages according to the virtual attributes of messages. Thus, we design two algorithms, submodular-based approximation algorithm and simulated annealing-based algorithm, to solve different scenarios of the problem. Both experiment and simulation results show that VITA can reduce the total traffic amount of redundant messages by 45%-75% and reduce the control overhead by 33%-80% compared with state-of-the-art solutions.

Energy and cost are two primary concerns when leveraging drones for urban sensing. With the advances of wireless charging technologies and the inspiration from the sparse crowdsensing paradigm, this paper proposes a novel drone-based collaborative sparse-sensing framework DroneSense, demonstrating its feasibility for sustainable urban-scale sensing. We focus on a typical use case, i.e., leveraging DroneSense to sense open parking spaces. DroneSense selects a minimum number of Points of Interest (POIs) to schedule drones for physical data sensing and then infers the parking occupancy of the remaining POIs to meet the overall quality requirement. However, drone-based sensing is different from human-centric crowdsensing, resulting in a series of new problems, including which POIs are visited first, when and where to charge drones, which drones to charge first, how much to charge, and when to stop the scheduling. To this end, we design a holistic solution, including context-aware matrix factorization for parking occupancy data inference, progressive determination of task quantity, deep reinforcement learning (DRL) based task selection, energy-aware DRL-based task scheduling, and adaptive charger scheduling. Extensive experiments with a real-world on-street parking dataset from Shenzhen, China demonstrate the obvious advantages of DroneSense.

Tag cloning and spoofing pose great challenges to RFID applications. This paper presents the design and evaluation of RCID, a novel system to fingerprint RFID tags based on the unique reflection coefficient of each tag circuit. Based on a novel OFDM-based fingerprint collector, our system can quickly acquire and verify each tag’s RCID fingerprint which are independent of the RFID reader and measurement environment. Our system applies to COTS RFID tags and readers after a firmware update at the reader. Extensive prototyped experiments on 600 tags confirm that RCID is highly secure with the authentication accuracy up to 97.15% and the median authentication error rate equal to 1.49%. RCID is also highly usable because it only takes about 8 s to enroll a tag and 2 ms to verify an RCID fingerprint with a fully connected multi-class neural network. Finally, empirical studies demonstrate that the entropy of an RCID fingerprint is about 202 bits over a bandwidth of 20 MHz in contrast to the best prior result of 17 bits, thus offering strong theoretical resilience to RFID cloning and spoofing.

Many efforts have been devoted to online pricing mechanism design for different settings. In this work, we consider a common but challenging setting where the buyers have private time-sensitive valuations and the seller has limited supply. The seller offers a take-it-or-leave-it posted price for each arriving buyer and aims to maximize the expected total revenue. The unknown distribution of time-sensitive valuations and limited supply significantly increase the difficulty of searching the optimal dynamic posted prices. Given B identical items to sell, when the time-dependent valuations can be estimated with a factor of α, we prove Ω(log(1/α)) lower bound with respect to the optimal fixed distribution over prices and design an algorithm achieving tight O(log(1/α)) competitive ratio. When the seller has no information about the future trends of buyers’ valuations, we prove Ω(log B) lower bound and show that there is an algorithm with tight O(log B) competitive ratio by modeling the problem as adversarial bandits with knapsacks optimization. Extensive simulation studies show that our algorithm outperforms previous mechanisms in various settings.

With the advancement of deep learning techniques, major cloud providers and niche machine learning service providers start to offer their cloud-based machine learning tools, also known as machine learning as a service (MLaaS), to the public. According to our measurement, for the same task, these MLaaSes from different providers have varying performance due to the proprietary datasets, models, etc. Federating different MLaaSes together allows us to improve the analytic performance further. However, naively aggregating results from different MLaaSes not only incurs significant momentary cost but also may lead to sub-optimal performance gain due to the introduction of possible false-positive results. In this paper, we propose Armol, a framework to federate the right selection of MLaaS providers to achieve the best possible analytic performance. We first design a word grouping algorithm to unify the output labels across different providers. We then present a deep combinatorial reinforcement learning based-approach to maximize the accuracy while minimizing the cost. The predictions from the selected providers are then aggregated together using carefully chosen ensemble strategies. The real-world trace-driven evaluation further demonstrates that Armol is able to achieve the same accuracy results with 67% less inference cost.

Building a fault-tolerant edge system that can quickly react to node overloads or failures is challenging due to the unreliability of edge devices and the strict service deadlines of modern applications. Moreover, unnecessary task migrations can stress the system network, giving rise to the need for a smart and parsimonious failure recovery scheme. Prior approaches often fail to adapt to highly volatile workloads or accurately detect and diagnose faults for optimal remediation. There is thus a need for a robust and proactive fault-tolerance mechanism to meet service level objectives. In this work, we propose PreGAN, a composite AI model using a Generative Adversarial Network (GAN) to predict preemptive migration decisions for proactive fault-tolerance in containerized edge deployments. PreGAN uses co-simulations in tandem with a GAN to learn a few-shot anomaly classifier and proactively predict migration decisions for reliable computing. Extensive experiments on a Raspberry-Pi based edge environment show that PreGAN can outperform state-of-the-art baseline methods in fault-detection, diagnosis and classification, thus achieving high quality of service. PreGAN accomplishes this by 5.1% more accurate fault detection, higher diagnosis scores and 23.8% lower overheads compared to the best method among the considered baselines.

In data center networks (DCNs), host-side packet processing accounts for a large portion of the end-to-end latency of TCP flows. Thus, the profiling of host-side latency anomalies has been considered as a crucial part in DCN performance diagnosis and troubleshooting. In particular, such profiling requires full coverage (i.e., profiling every TCP packet handled by end-hosts) and low overhead (i.e., profiling should avoid high CPU consumption in end-hosts). However, existing solutions fully rely on end-hosts to implement host-side latency profiling, leading to low coverage or high overhead. In this paper, we propose Torp, a framework that offers full-coverage and low-overhead profiling of host-side latency. Our key idea is to offload profiling operations to top-of-rack (ToR) switches, which inherently offer full coverage and line-rate packet processing performance. Specifically, Torp selectively offloads profiling operations to the ToR switch based on switch limitations. It efficiently coordinates the ToR switch and end-hosts to execute the entire latency profiling task. We have implemented Torp on 32×100Gbps Tofino switches. Testbed experiments indicate that Torp achieves full coverage and orders of magnitude lower host-side overhead compared to other solutions.

Today’s network switches often use on-chip shared memory to improve buffer efficiency and absorb bursty traffic. Current buffer management practices usually rely on simple heuristics and have unrealistic assumptions about the traffic pattern, since developing a buffer management policy suited for every scenario is infeasible. We show that modern machine learning techniques can be of essential help to learn efficient policies automatically.In this paper, we propose Neural Dynamic Threshold (NDT) that uses deep reinforcement learning (RL) to learn buffer management policies without human instructions except for a high-level objective. To tackle the high complexity and scale of the buffer management problem, we develop two domain-specific techniques upon off-the-shelf deep RL solutions. First, we design a scalable RL model by leveraging the permutation symmetry of the switch ports. Second, we use a two-level control mechanism to achieve efficient training and decision-making. The buffer allocation is directly controlled by a low-level heuristic during the decision interval, while the RL agent only decides the high-level control factor according to the traffic density. Testbed and simulation experiments demonstrate that NDT generalizes well and outperforms hand-tuned heuristic policies even on workloads for which it was not explicitly trained.

Long Range Wide Area Network (LoRaWAN), using the linear chirp for data modulation, is known for its low-power and long-distance communication to connect massive Internet-of-Things devices at a low cost. However, LoRaWAN throughput is far behind the demand for the dense and large-scale IoT deployments, due to the frequent collisions with the by-default random channel access (i.e., ALOHA). Recently, some works enable an effective LoRa carrier-sense for collision avoidance. However, the continuous back-off makes the network throughput easily saturated and degrades the energy efficiency at LoRa end nodes. In this paper, we propose CurveALOHA, a brand-new media access control scheme to enhance the throughput of random channel access by embracing non-linear chirps enabled quasi-orthogonal logical channels. First, we empirically show that non-linear chirps can achieve similar noise tolerance ability as the linear one does. Then, we observe that multiple nonlinear chirps can create new logical channels which are quasi-orthogonal with the linear one and each other. Finally, given a set of non-linear chirps, we design two random chirp selection methods to guarantee an end node can access a channel with less collision probability. We implement CurveALOHA with the software-defined radios and conduct extensive experiments in both indoor and outdoor environments. The results show that CurveALOHA’s network throughput is 59.6% higher than the state-of-the-art carrier-sense MAC.

To improve the utility of learning applications and render machine learning solutions feasible for complex applications, a substantial amount of heavy computations is needed. Thus, it is essential to delegate the computations among several workers, which brings up the major challenge of coping with delays and failures caused by the system’s heterogeneity and uncertainties. In particular, minimizing the end-to-end job in-order execution delay, from arrival to delivery, is of great importance for real-world delay-sensitive applications. In this paper, for computation of each job iteration in a stochastic heterogeneous distributed system where the workers vary in their computing and communicating powers, we present a novel joint scheduling-coding framework that optimally split the coded computational load among the workers. This closes the gap between the workers’ response time, and is critical to maximize the resource utilization. To further reduce the in-order execution delay, we also incorporate redundant computations in each iteration of a distributed computational job. Our simulation results demonstrate that the delay obtained using the proposed solution is dramatically lower than the uniform split which is oblivious to the system’s heterogeneity and, in fact, is very close to an ideal lower bound just by introducing a small percentage of redundant computations.

A promising approach for leveraging the flexibility and mitigating the complexity of future telecom systems is the use of machine learning (ML) models that can analyze the network performance, as well as taking proactive actions. A key enabler for ML models is timely access to reliable data, in terms of features, which require pervasive measurement points throughout the network. However, excessive monitoring is associated with network overhead. Considering domain knowledge may provide clues to find a balance between overhead reduction and meeting requirements on future ML use cases by monitoring just enough features. In this work, we propose a method of unsupervised feature selection that provides a structured approach in incorporation of the domain knowledge in terms of policies. Policies are provided to the method in form of must-have features defined as the features that need to be monitored at all times. We name such family of unsupervised feature selection the policy-induced unsupervised feature selection as the policies inform selection of the latent features. We evaluate the performance of the method on two rich sets of data traces collected from a data center and a 5G-mmWave testbed. Our empirical evaluations point at the effectiveness of the solution.

The rapid upgrade of link speed and the prosperity of new applications in data center networks (DCNs) lead to a rigorous demand for ultra-low latency and high throughput. To mitigate the overhead of traditional software-based packet processing at end-hosts, RDMA (Remote Direct Memory Access) has been widely adopted in DCNs. Particularly, congestion control (CC) mechanisms designed for RDMA have attracted much attention to avoid performance deterioration when packets lose. However, through comprehensive analysis, we found that existing RDMA CC schemes have limitations of a sluggish response to congestion and unawareness of tiny microbursts due to the long end-to-end control loop. In this paper, we propose PACC, a switch-driven RDMA CC algorithm with easy deployability. PACC is driven by PI controller-based computation, threshold-based flow discrimination and weight-based allocation at the switch. It leverages real-time queue length to generate accurate congestion feedback proactively and piggybacks it to the corresponding source without modification to end-hosts. We theoretically analyze the stability and key parameter settings of PACC. Then, we conduct both micro-benchmark and large-scale simulations to evaluate the performance of PACC. The results show that PACC achieves fairness, fast reaction, high throughput, and 6~69% lower FCT (Flow Completion Time) than DCQCN, TIMELY and HPCC.

State-of-the-art blockchain sharding solutions, say Monoxide, can induce imbalanced transaction (TX) distributions among all blockchain shards due to their account deployment mechanisms. Imbalanced TX distributions then cause hot shards, in which the cross-shard TXs may experience an unlimited length of confirmation latency. Thus, how to address the hot-shard issue and how to reduce cross-shard TXs become significant challenges of blockchain state sharding. Through reviewing the related studies, we find that a cross-shard TX protocol that can achieve workload balance among all shards and simultaneously reduce the number of cross-shard TXs is still absent from the literature. To this end, we propose BrokerChain, which is a cross-shard blockchain protocol devised for the account/balance-based state sharding. Essentially, BrokerChain exploits fine-grained state partition and account segmentation. We also elaborate on how BrokerChain handles cross-shard TXs through broker accounts. The security issues and other properties of BrokerChain are analyzed substantially. Finally, we conduct comprehensive evaluations using both a cloud-based prototype and a transaction-driven simulator. The evaluation results show that BrokerChain outperforms other solutions in terms of system throughput, transaction confirmation latency, the queue size of transaction pool, and workload balance.

To conduct service quality management of industry devices or Internet infrastructures, various deep learning approaches have been used for extracting the normal patterns of multivariate Key Performance Indicators (KPIs) for unsupervised anomaly detection. However, in the scenario of Content Delivery Networks (CDN), KPIs that belong to diverse websites usually exhibit various structures at different timesteps and show the non-stationary sequential relationship between them, which is extremely difficult for the existing deep learning approaches to characterize and identify anomalies. To address this issue, we propose a switching Gaussian mixture variational recurrent neural network (SGmVRNN) suitable for multivariate CDN KPIs. Specifically, SGmVRNN introduces the variational recurrent structure and assigns its latent variables into a mixture Gaussian distribution to model complex KPI time series and capture the diversely structural and dynamical characteristics within them, while in the next step it incorporates a switching mechanism to characterize these diversities, thus learning richer representations of KPIs. For efficient inference, we develop an upward-downward autoencoding inference method which combines the bottom-up likelihood and up-bottom prior information of the parameters for accurate posterior approximation. Extensive experiments on real-world data show that SGmVRNN significantly outperforms the state-of-the-art approaches according to F1-score on CDN KPIs from diverse websites.