Latency Constraints Research Articles

Future of Drug Discovery: The Synergy of Edge Computing, Internet of Medical Things, and Deep Learning

The global spread of COVID-19 highlights the urgency of quickly finding drugs and vaccines and suggests that similar challenges will arise in the future. This underscores the need for ongoing efforts to overcome the obstacles involved in the development of potential treatments. Although some progress has been made in the use of Artificial Intelligence (AI) in drug discovery, virologists, pharmaceutical companies, and investors seek more long-term solutions and greater investment in emerging technologies. One potential solution to aid in the drug-development process is to combine the capabilities of the Internet of Medical Things (IoMT), edge computing (EC), and deep learning (DL). Some practical frameworks and techniques utilizing EC, IoMT, and DL have been proposed for the monitoring and tracking of infected individuals or high-risk areas. However, these technologies have not been widely utilized in drug clinical trials. Given the time-consuming nature of traditional drug- and vaccine-development methods, there is a need for a new AI-based platform that can revolutionize the industry. One approach involves utilizing smartphones equipped with medical sensors to collect and transmit real-time physiological and healthcare information on clinical-trial participants to the nearest edge nodes (EN). This allows the verification of a vast amount of medical data for a large number of individuals in a short time frame, without the restrictions of latency, bandwidth, or security constraints. The collected information can be monitored by physicians and researchers to assess a vaccine’s performance.

Q-Learning-Aided Offloading Strategy in Edge-Assisted Federated Learning over Industrial IoT

Federated learning (FL) is a key solution to realizing a cost-efficient and intelligent Industrial Internet of Things (IIoT). To improve training efficiency and mitigate the straggler effect of FL, this paper investigates an edge-assisted FL framework over an IIoT system by combining it with a mobile edge computing (MEC) technique. In the proposed edge-assisted FL framework, each IIoT device with weak computation capacity can offload partial local data to an edge server with strong computing power for edge training. In order to obtain the optimal offloading strategy, we formulate an FL loss function minimization problem under the latency constraint in the proposed edge-assisted FL framework by optimizing the offloading data size of each device. An optimal offloading strategy is first derived in a perfect channel state information (CSI) scenario. Then, we extend the strategy into an imperfect CSI scenario and accordingly propose a Q-learning-aided offloading strategy. Finally, our simulation results show that our proposed Q-learning-based offloading strategy can improve FL test accuracy by about 4.7% compared to the conventional FL scheme. Furthermore, the proposed Q-learning-based offloading strategy can achieve similar performance to the optimal offloading strategy and always outperforms the conventional FL scheme in different system parameters, which validates the effectiveness of the proposed edge-assisted framework and Q-learning-based offloading strategy.

Online Learning for Rate-Adaptive Task Offloading Under Latency Constraints in Serverless Edge Computing

We consider the interplay between latency constrained applications and function-level resource management in a serverless edge computing environment. We develop a game theoretic model of the interaction between rate adaptive applications and a load balancing operator under a function-oriented pay-as-you-go pricing model. We show that under perfect information, the strategic interaction between the applications can be formulated as a generalized Nash equilibrium problem, and use variational inequality theory to prove that the game admits an equilibrium. For the case of imperfect information, we propose an online learning algorithm for applications to maximize their utility through rate adaptation and resource reservation. We show that the proposed algorithm can converge to equilibria and achieves zero regret asymptotically, and our simulation results show that the algorithm achieves good system performance at equilibrium, ensures fast convergence, and enables applications to meet their latency constraints.

UAV-Enabled Mobile-Edge Computing for AI Applications: Joint Model Decision, Resource Allocation, and Trajectory Optimization

Due to the flexible mobility and agility, unmanned aerial vehicles (UAVs) are expected to be deployed as aerial base stations (BSs) in future air-ground integrated wireless networks, providing temporary and controllable coverage and additional computation capabilities for ground Internet of Things (IoT) devices with or without infrastructure support. Meanwhile, with the breakthrough of artificial intelligence (AI), more and more AI applications relying on AI methods such as deep neural networks (DNNs) are expected to be applied in various fields such as smart homes, smart factories and smart cities to improve our lifestyles and efficiency dramatically. However, AI applications are generally computation-intensive, latency-sensitive, and energy-consuming, making resource-constrained IoT devices unable to benefit from AI anytime and anywhere. In this paper, we study mobile edge computing (MEC) for AI applications in air-ground integrated wireless networks. Our goal is to minimize the service latency while ensuring the learning accuracy requirements and energy consumption. To achieve that, we take DNN as the typical AI application and formulate an optimization problem that optimizes the DNN model decision, computation and communication resource allocation, and UAV trajectory control, subject to the energy consumption, latency, computation and communication resource constraints. Considering the formulated problem is non-convex, we decompose it into multiple convex subproblems and then alternately solve them till they converge to the desired solution. Simulation results show that the proposed algorithm significantly improves the system performance for AI applications.

Cost optimization of omnidirectional offloading in two-tier cloud–edge federated systems

The use of the federation can exploit the advantages of cloud and edge computing technologies as the federation provides the facilities by which both can complement each other. However, certain services are needed to offload from clouds to edges, termed reverse offloading, and between edges, termed horizontal offloading. By considering the scenarios discussed above, in this paper, we propose a generic omnidirectional (OMNI) architecture of cloud–edge computing systems intending to provide vertical (edge to cloud offloading, and vice versa), and horizontal (between edges) offloading. To investigate the effectiveness of the proposed architecture in different operational scenarios, we formulate the cost optimization problem with different latency (loose, low, ultra-low) constraints. We develop an offloading algorithm using simulated annealing named the two-tier simulated annealing (TTSA) algorithm. We set two criteria for offloading: (1) offloading based on the job’s size and (2) offloading based on the job’s priority, irrespective of its size. The experimental results show that our proposed OMNI architecture can reduce the total cost by 7%–10%, compared to other existing architectures, and our proposed TTSA algorithm can reduce the cost by 45%–55%, compared to other existing algorithms. The average latency in OMNI architecture is relatively very less compared to other architectures.

Computation Offloading and Resource Allocation in Unmanned Aerial Vehicle Networks

Deploying unmanned aerial vehicles (UAVs) as mobile edge computing (MEC) servers has been attracting significant attention, since the UAVs' inherent maneuverability and mobility can further reduce the distance between the users and computational functionalities. In the UAV-assisted MEC systems, the caching and offloading decision optimization, subject to the computation and energy constraints of the UAVs, is one of the key design issues. Against this backdrop, in this paper, we propose a novel cloud-edge framework to facilitate MEC in the UAV networks. Specifically, in our framework, the edge UAVs (EUAVs), together with the cloud, provide caching and computing services for the terrestrial users. In order to minimize the weighted sum cost of latency and energy consumption, we jointly optimize the caching and offloading decisions, the EUAV deployment, the radio and computation resource allocation, while simultaneously satisfying the UAVs' cache and computation capacity constraints, as well as the users' latency and energy consumption constraints. To solve this NP-hard problem, we propose a sequential convex programming (SCP) and sequential quadratic programming (SQP) based deep Q-learning (SS-DQN) algorithm. The proposed algorithm allows the system to adaptively adjust caching and offloading decisions with the EUAV deployment scheme and the resource allocation scheme obtained by SCP and SQP algorithms, respectively. Simulation results verify the superiority of our proposed algorithm compared to two benchmark schemes, and demonstrate how the combination of DRL and convex optimization can improve the system performance significantly.

Real-Time HAP-Assisted Vehicular Edge Computing for Rural Areas

Non-Terrestrial Networks (NTNs) are expected to be a key component of 6th generation (6G) networks to support broadband seamless Internet connectivity and expand the coverage even in rural and remote areas. In this context, High Altitude Platforms (HAPs) can act as edge servers to process computational tasks offloaded by energy-constrained terrestrial devices such as Internet of Things (IoT) sensors and ground vehicles (GVs). In this letter, we analyze the opportunity to support Vehicular Edge Computing (VEC) via HAP in a rural scenario where GVs can decide whether to process data onboard or offload them to a HAP. We characterize the system as a set of queues in which computational tasks arrive according to a Poisson arrival process. Then, we assess the optimal VEC offloading factor to maximize the probability of real-time service, given latency and computational capacity constraints.

Model-driven Cluster Resource Management for AI Workloads in Edge Clouds

Since emerging edge applications such as Internet of Things (IoT) analytics and augmented reality have tight latency constraints, hardware AI accelerators have been recently proposed to speed up deep neural network (DNN) inference run by these applications. Resource-constrained edge servers and accelerators tend to be multiplexed across multiple IoT applications, introducing the potential for performance interference between latency-sensitive workloads. In this article, we design analytic models to capture the performance of DNN inference workloads on shared edge accelerators, such as GPU and edgeTPU, under different multiplexing and concurrency behaviors. After validating our models using extensive experiments, we use them to design various cluster resource management algorithms to intelligently manage multiple applications on edge accelerators while respecting their latency constraints. We implement a prototype of our system in Kubernetes and show that our system can host 2.3× more DNN applications in heterogeneous multi-tenant edge clusters with no latency violations when compared to traditional knapsack hosting algorithms.

Energy-Aware and Secure Task Offloading for Multi-Tier Edge-Cloud Computing Systems

Nowadays, Unmanned Aerial Vehicle (UAV) devices and their services and applications are gaining popularity and attracting considerable attention in different fields of our daily life. Nevertheless, most of these applications and services require more powerful computational resources and energy, and their limited battery capacity and processing power make it difficult to run them on a single device. Edge-Cloud Computing (ECC) is emerging as a new paradigm to cope with the challenges of these applications, which moves computing resources to the edge of the network and remote cloud, thereby alleviating the overhead through task offloading. Even though ECC offers substantial benefits for these devices, the limited bandwidth condition in the case of simultaneous offloading via the same channel with increasing data transmission of these applications has not been adequately addressed. Moreover, protecting the data through transmission remains a significant concern that still needs to be addressed. Therefore, in this paper, to bypass the limited bandwidth and address the potential security threats challenge, a new compression, security, and energy-aware task offloading framework is proposed for the ECC system environment. Specifically, we first introduce an efficient layer of compression to smartly reduce the transmission data over the channel. In addition, to address the security issue, a new layer of security based on an Advanced Encryption Standard (AES) cryptographic technique is presented to protect offloaded and sensitive data from different vulnerabilities. Subsequently, task offloading, data compression, and security are jointly formulated as a mixed integer problem whose objective is to reduce the overall energy of the system under latency constraints. Finally, simulation results reveal that our model is scalable and can cause a significant reduction in energy consumption (i.e., 19%, 18%, 21%, 14.5%, 13.1% and 12%) with respect to other benchmarks (i.e., local, edge, cloud and further benchmark models).

Scheduling IDK classifiers with arbitrary dependences to minimize the expected time to successful classification

This paper introduces and evaluates a general construct for trading off accuracy and overall execution duration in classification-based machine perception problems—namely, the generalized IDK classifier cascade. The aim is to select the optimal sequence of classifiers required to minimize the expected (i.e. average) execution duration needed to achieve successful classification, subject to a constraint on quality, and optionally a latency constraint on the worst-case execution duration. An IDK classifier is a software component that attempts to categorize each input provided to it into one of a fixed set of classes, returning “I Don’t Know” (IDK) if it is unable to do so with the required level of confidence. An ensemble of several different IDK classifiers may be available for the same classification problem, offering different trade-offs between effectiveness (i.e. the probability of successful classification) and timeliness (i.e. execution duration). A model for representing such characteristics is defined, and a method is proposed for determining the values of the model parameters for a given ensemble of IDK classifiers. Optimal algorithms are developed for sequentially ordering IDK classifiers into an IDK cascade, such that the expected duration to successfully classify an input is minimized, optionally subject to a latency constraint on the worst-case overall execution duration of the IDK cascade. The entire methodology is applied to two real-world case studies. In contrast to prior work, the methodology developed in this paper caters for arbitrary dependences between the probabilities of successful classification for different IDK classifiers. Effective practical solutions are developed considering both single and multiple processors.

A new clustering‐based optimised energy approach for fog‐enabled IoT networks

AbstractThe Internet of Things (IoT) has induced many advances in the modern world, thus always arousing great interest from the scientific community. Due to a large number of connected devices, this paradigm has put into circulation an enormous data quantity to be processed and offloaded while respecting latency and energy constraints that Central Cloud alone cannot meet. Hence, Fog Computing has come to fill these gaps by providing computing, management, and storage resources via small‐distributed data centres located at the network edge. The majority of previous research on centralised clustering in an IoT environment did not consider both residual energy and the location of the IoT devices, as well as the signalling communication frequency, to solve an issue of extreme energy consumption during data exchange between IoT devices and Fog entities. Thus, the objective of this study is to find solutions taking into account these criteria to guarantee the Fog system's energy efficiency while coping with the energy consumption increase related to the handling of massive data volume. Indeed, the proposed study presents a novel approach based on Fog‐IoT architecture to ensure intelligent energy management during communication and offloaded task processing in IoT applications. The simulation results show its effectiveness.

Approximation Opportunities in Edge Computing Hardware: A Systematic Literature Review

With the increasing popularity of the Internet of Things and massive Machine Type Communication technologies, the number of connected devices is rising. However, although enabling valuable effects to our lives, bandwidth and latency constraints challenge Cloud processing of their associated data amounts. A promising solution to these challenges is the combination of Edge and approximate computing techniques that allows for data processing nearer to the user. This article aims to survey the potential benefits of these paradigms’ intersection. We provide a state-of-the-art review of circuit-level and architecture-level hardware techniques and popular applications. We also outline essential future research directions.

Deep Reinforcement Learning-Based Grant-Free NOMA Optimization for mURLLC

Grant-free non-orthogonal multiple access (GF-NOMA) is a potential technique to support massive Ultra-Reliable and Low-Latency Communication (mURLLC) service. However, the dynamic resource configuration in GF-NOMA systems is challenging due to random traffics and collisions, that are unknown at the base station (BS). Meanwhile, joint consideration of the latency and reliability requirements makes the resource configuration of GF-NOMA for mURLLC more complex. To address this problem, we develop a novel learning framework for signature-based GF-NOMA in mURLLC service taking into account the multiple access signature collision, the UE detection, as well as the data decoding procedures for the K-repetition GF and the Proactive GF schemes. The goal of our learning framework is to maximize the long-term average number of successfully served users (UEs) under the latency constraint. We first perform a real-time repetition value configuration based on a double deep Q-Network (DDQN) and then propose a Cooperative Multi-Agent learning technique based DQN (CMA-DQN) to optimize the configuration of both the repetition values and the contention-transmission unit (CTU) numbers. Our results show the superior performance of CMA-DQN over the conventional load estimation-based uplink resource configuration approach (LE-URC) in heavy traffic and demonstrate its capability in dynamically configuring in long term for mURLLC service. In addition, with our learning optimization, the Proactive scheme always outperforms the K-repetition scheme in terms of the number of successfully served UEs, especially under the high backlog traffic scenario.

Deep Learning at the Physical Layer for Adaptive Terahertz Communications

Wireless communications in the terahertz (THz) band will become a cornerstone of sixth-generation (6G) networks. The THz channel, however, presents several challenges, such as distance-dependent absorption coefficients that can change the bandwidth significantly in case of mobility. Thus, future THz transmitters will have to switch modulation and bandwidth almost continuously. Moreover, using the same transmission scheme can enable adversaries to leverage smart interfering to inflict more damage with less energy expense. To help enable adaptive and secure THz communications, this article presents the first ever experimental study of modulation and bandwidth classification (MBC) at THz frequencies through deep learning (DL) techniques. We have performed an extensive experimental data collection campaign at 120 GHz with different modulation schemes, signal bandwidth (up to 20 GHz), and different signal-to-noise ratio (SNR) levels. We prove for the first time the feasibility and effectiveness of MBC at THz frequencies, with our DL models reaching accuracy up to 78% and 90% in low- and high-SNR conditions. Furthermore, we investigate the memory and latency constraints that need to be satisfied as a function of the signal bandwidth, and propose a boosting technique to improve the inference quality by trading off latency for accuracy. Finally, we experimentally evaluate the latency of our CNN models through FPGA implementation.

Joint Allocations of Radio and Computational Resource for User Energy Consumption Minimization Under Latency Constraints in Multi-Cell MEC Systems

This paper investigates the joint allocation of radio resources (i.e., base stations (BSs), sub-channels, and uplink transmission powers) and computational resources for a multicell mobile edge computing (MEC) system, aimed at minimizing the sum energy consumption of all mobile terminals under strict delay and signal-to-interference-plus-noise ratio (SINR) constraints. The problem is challenging to solve due to the highly coupled mixed-integer variables and the complicated expressions of the constraints. To tackle the difficulty, we decompose the problem into two sub-problems: the power and computational resource allocation sub-problem and the user-BS-subchannel association sub-problem. For the first sub-problem, we derive the optimal computational resource as a function of the optimal transmission power. Then the power optimization problem is distributed to each cell, and the local-optimal power solution within each cell is derived in closed-form expressions. For the second sub-problem, we design an optimal modified-cutting-plane (MCP) algorithm with exponential time-complexity. A polynomial-time near-optimal pivoting-and-subgradient (PS) algorithm is further proposed to reduce the complexity. Simulation results demonstrate that the PS algorithm achieves comparable performance to the optimal MCP algorithm, while they both outperform the benchmark schemes in most cases.

Revenue Maximizing Online Service Function Chain Deployment in Multi-Tier Computing Network

Multi-tier computing (MC) is a promising architecture that integrates cloud computing, fog computing, and edge computing to provide users with a consistent experience of computing services by fusing computing devices within the network through virtualization technology. Although MC combines powerful computation and communication resources, the massive demand from Service Function Chain (SFC) deployments continues to make it challenging regarding resource constraints, latency satisfaction, and revenue-cost tradeoffs. To this end, in this article, we study an SFC deployment problem in MC and formulate a problem for maximizing the revenue of online SFC deployment under latency, computation resources, and communication resources constraints. To solve this online problem better, we construct a computation and communication resource cost model and transform the original online problem into a deployment cost minimization problem and a request admission problem by an alternating optimization approach. To solve the two subproblems, we propose an online approximation algorithm with a provable competitive ratio for the particular scenario with no latency requirements. Then, based on the cost model, we propose an online heuristic algorithm that adopts a binary search method for the original problem with latency requirements. Simulation experiments show that our two proposed online algorithms have advantages in total revenue, running time, and load balancing compared with other comparison algorithms.

Energy-Efficient Joint Computation Offloading and Resource Allocation Strategy for ISAC-Aided 6G V2X Networks

Integrated sensing and communications (ISAC) is a pillar technology of 6G to enable intelligent environment awareness and hardware cost efficiencies. In ISAC-aided 6G vehicle-to-everything (V2X) networks, the perception data fusion from multiple sources can make a reliable and efficient data sharing to guarantee driving safety. However, the contradiction between task delay requirement and energy consumption becomes more and more prominent with the growth of computing task amount. Therefore, this paper proposes a joint computation offloading and resource allocation strategy to build greener V2X networks with mobile-edge computing (MEC) and ISAC technologies. A data fusion architecture for cooperative perception is first introduced to support fusing massive perception data from wireless infrastructures and vehicles. Then, a minimization problem of queuing latency is formulated with long-term latency and energy consumption constraints for data fusion computing tasks. The problem is further reformulated by the Lyapunov optimization method to transfer delay and energy constraints into queue stability problems. Finally, a joint computation offloading and resource allocation (JCORA) scheme is proposed to obtain the optimal computation offloading and resource allocation decision, achieving the balance between task delay and energy consumption. Extensive simulations validate the effectiveness of the proposed strategy compared with other baseline schemes.

Selective Polling and Controlled Contention Based WLAN MAC Scheme for Low-Latency Applications

A novel medium access control scheme is proposed and investigated for central coordinated wireless local area networks to support future low-latency applications. The scheme divides each beacon interval into multiple service cycles, and each service cycle is composed of a polling period (PP) and a contention period (CP). By polling stations according to their latest channel accessing times during each PP and properly assigning backoff count for each station during each CP, the proposed scheme reduces the resource waste caused by polling empty stations and collisions, thereby achieving low-latency performance. In each service cycle, the number of stations to poll during the PP is calculated based on station empty probability, and the backoff count assigned to a station that has not been polled during the PP is unique and relatively small. Our simulation results show that the proposed scheme reduces uplink latency effectively, and can afford more traffic under the 0.5 ms uplink latency constraint compared to the conventional schemes.

DRL-assisted delay optimized task offloading in automotive-industry 5.0 based VECNs

The rapid growth of Automotive-Industry 5.0 and its emergence with beyond fifth-generation (B5G) communications, is making vehicular edge computing networks (VECNs) increasingly complex. The latency constraints of modern automotive applications make it difficult to run complex applications on vehicle on-board units (OBUs). While multi-access edge computing (MEC) can facilitate task offloading to execute these applications, it is still a challenge to access them promptly and optimally. Traditional algorithms struggle to guarantee accuracy in such dynamic environment, but deep reinforcement learning (DRL) methods offer improved accuracy, robustness, and real-time decision-making capabilities. In this paper, we propose a DRL-based mobility, contact, and load aware cooperative task offloading (DCTO) scheme. DCTO is designed for both cellular and mmWave radio access technologies (RATs), and both binary and partial offloading mechanisms. DCTO targets delay minimization by opportunistically switching RATs and offloading mechanisms. We consider relative efficacy and neutrality factors as key performance indicators and use them to derive the DRL agent’s reward function. Extensive evaluations demonstrate that the DCTO scheme exhibits a substantial enhancement in task success rate, with an increase from 2.61% to 21.34%. It also improves the efficacy factor from 1.38 to 3.52 and reduces the neutrality factor from 4.99 to 0.76. Furthermore, the average task processing time is reduced by a range of 3.77% to 24.15%. Additionally, the DCTO scheme outperforms the other evaluated schemes in terms of reward and TFPS ratio.

WatchDog: Real-time Vehicle Tracking on Geo-distributed Edge Nodes

Vehicle tracking, a core application to smart city video analytics, is becoming more widely deployed than ever before thanks to the increasing number of traffic cameras and recent advances in computer vision and machine-learning. Due to the constraints of bandwidth, latency, and privacy concerns, tracking tasks are more preferable to run on edge devices sitting close to the cameras. However, edge devices are provisioned with a fixed amount of computing budget, making them incompetent to adapt to time-varying and imbalanced tracking workloads caused by traffic dynamics. In coping with this challenge, we propose WatchDog, a real-time vehicle tracking system that fully utilizes edge nodes across the road network. WatchDog leverages computer vision tasks with different resource-accuracy tradeoffs, and decomposes and schedules tracking tasks judiciously across edge devices based on the current workload to maximize the number of tasks while ensuring a provable response time-bound at each edge device. Extensive evaluations have been conducted using real-world city-wide vehicle trajectory datasets, achieving exceptional tracking performance with a real-time guarantee.