Main Conference
Track 4 – Emerging Technologies, Architectures and Services
Session T4-S6
Edge Computing and Caching
Conference
11:00 AM
—
12:30 PM KST
Local
May 26 Tue, 9:00 PM
—
10:30 PM CDT
Overlay Coded Multicast for Edge Caching in 5G-Satellite Integrated Networks
Xinmu Wang, Hewu Li, Tianming Lan and Qian Wu (Tsinghua University, China)
0
Edge caching in 5G networks shortens latency and alleviates the backhaul. Bringing contents from the core network to the caches is a critical issue. When the satellite is integrated with the terrestrial 5G system, an overlay can be formed for cost-efficient content delivery. Multicast delivery over satellite is a promising scheme due to the broadcast nature of the wireless medium and wide coverage. Requests for popular content at nearby times can be aggregated through a multicast stream for bandwidth efficiency. However, directly multicast to a large audience from the satellite suffers from the problems led by fading channels and the flat topologies. The overlay architecture provides new solutions to handle the drawbacks of satellite multicast, e.g., channel and reception diversity, the difficulty of loss recovery and feedback explosion. We explicitly introduce the overlay architecture based on the configuration of multiple multicast groups and the merging of base station clusters for each communication session. The operation of this overlay is further explained. Besides, we also apply network coding to the multicast and cache networks to improve data recovery and bandwidth efficiency. Both theoretical analysis and numerical experiments demonstrate the optimization of network performance.
An EPEC Analysis among Mobile Edge Caching, Content Delivery Network and Data Center
Yue Yu (Southeast University, China); Xiao Tang (Northwestern Polytechnical University, China); Yiyong Zha and Yunfei Zhang (Tencent, China); Tiecheng Song (National Mobile Communications Research Laboratory, Southeast University, China); Zhu Han (University of Houston, USA)
0
Mobile edge caching (MEC), content delivery network (CDN) and data center (DC) serve Internet content providers (ICPs) with different advantages and disadvantages. In this paper, we propose an equilibrium problem with equilibrium constraints (EPEC) to investigate the delivery strategies for files and pricing mechanisms for MEC, CDN, and DC. At the upper level, MEC and CDN predict the files' rational delivery strategies and set the delivery price for each byte to provide the content delivery service. DC serves as origin servers to provide free content delivery service. At the lower level, the files observe the price strategy and determine their delivery strategies. In the proposed EPEC problem, there exist Nash equilibriums, which are coupled with each other, at both the upper level and lower level. We adopt a block coordinate descent (BCD) method to find the equilibrium solutions at both the upper level and lower level. Simulation results show that our proposed approach yields high utilities at the equilibrium.
Design and Implementation on a LoRa System with Edge Computing
Zhiming Liu, Qihao Zhou and Lu Hou (Beijing University of Posts and Telecommunications, China); Rongtao Xu (Beijing Jiaotong University, China); Kan Zheng (Beijing University of Posts&Telecommunications, China)
0
The Long Range (LoRa) systems usually process all the computing tasks on the LoRa central server remotely, which brings large latency to Internet of Things (IoT) applications. In this paper, we propose a new design of a LoRa system with edge computing at the LoRa gateway. Our design enables that some of the time computing tasks for latency-sensitive applications can be dealt with timely. The implementation details of the LoRa gateway are presented along with functionality of each component. Finally, comprehensive experiments are conducted to evaluate the performance of the proposed system. The results show that the proposed system can decrease the latency of IoT applications and balance the workloads between the LoRa central server and the LoRa gateway.
Collaborative Edge Computing and Caching in Vehicular Networks
Zhuoxing Qin and Supeng Leng (University of Electronic Science and Technology of China, China); Jihua Zhou (Institute of Computing Technology, Chinese Academy of Sciences, China); Sun Mao (University of Electronic Science and Technology of China, China)
0
Mobile Edge Computing (MEC) can significantly promote the development of Internet of Vehicles (IoV) for providing a low-latency and high-reliability environment. Nevertheless, a huge amount of sensor data or computing requirements generated by massive vehicles in adjacent area may be duplicated. In order to realize the efficient diffusion of information, we propose a hierarchical end-edge framework with the aid of deep collaboration among data communication, computation offloading and content caching to minimize network overheads. Specially, duplicated perceived data and computation results are cached in advance to decrease repeated data uploading and duplicated computation in offloading process. In addition, the problem is formulated as a mixed integer non-linear programming (MINLP) problem, and the deep deterministic policy gradient (DDPG)-based resource allocation scheme is utilized to obtain a sub-optimal solution with low computation complexity. Performance evaluation demonstrates that the proposed scheme can significantly reduce network overheads compared with other benchmark methods.
Latency Guaranteed Edge Inference via Dynamic Compression Ratio Selection
Xiufeng Huang and Sheng Zhou (Tsinghua University, China)
0
With the development of intelligent Internet of things (IoT) devices, implementing machine learning algorithms at the network edge has become essential to many applications, such as autonomous driving, environment monitoring. However, the limited computation capability and energy constraint results in difficulties of running complex machine learning algorithms on edge devices subject to latency requirements, and one solution is to offload the computation tasks to the edge server. However, the wireless transmission of raw data from devices to the server is time consuming and may violate the latency requirement. To this end, lossy data compression can be helpful, but the information loss may lead to erroneous learning result, e.g., wrong classification. In this paper, we propose a transmission scheme with compression ratio selection for inference tasks with task completion latency guarantee. By dynamically selecting the optimal compression ratio with the awareness of the remaining latency budget, more tasks can be timely completed and get the correct inference results under the communication resource constraint. Furthermore, retransmitting less compressed data of tasks with erroneous inference results can potentially enhance the average accuracy. However, it is often hard to know whether the inference result is correct or not. We therefore use uncertainty to estimate the confidence of the results, and based on that, jointly optimize the retransmission and compression ratio selection.
Session Chair
Youngbin Im (UNIST, Korea)
Session T4-S7
Learning for Networks
Conference
2:00 PM
—
3:30 PM KST
Local
May 27 Wed, 12:00 AM
—
1:30 AM CDT
QLACO: Q-learning Aided Ant Colony Routing Protocol for Underwater Acoustic Sensor Networks
Zhengru Fang, Jingjing Wang and Chunxiao Jiang (Tsinghua University, Beijing, China); Biling Zhang (Beijing University of Posts and Telecommunications, China); Chuan Qin (Tsinghua University, China); Yong Ren (Tsinghua University, Beijing, China)
0
Recently, the technology of underwater wireless sensors networks (UWSNs) has received more attention on the exploitation of marine resources. However, underwater acoustic communication is still the only reliable means of ocean communication, which is entirely different from the terrestrial scene. In this paper, we propose Q-learning aided ant colony routing protocol (QLACO) to address the issues of energy-efficiency and link instability in UWSNs, which uses both the reward mechanism and artificial ants to determine a global optimal routing selection. QLACO uses the reward function to adapt to the dynamic underwater environment and enhance the packet delivery ratio (PDR). Moreover, we propose an anti-void mechanism to solve the void region dilemma. Simulation results show that QLACO outperforms Q-learning-based energy-efficient and lifetime-aware routing protocol (QELAR) and the depth-based protocol (DBR) in terms of PDR, energy consumption and latency.
DeepCReg: Improving Cellular-based Outdoor Localization using CNN-based Regressors
Karim Elawaad, Mohamed Ezzeldin and Marwan Torki (Alexandria University, Egypt)
0
In this paper, we propose DeepCReg, a convolutional neural network based regressor, that leverages the ubiquitous cellular data to estimate the location of the user in an outdoor environment. We formulate the problem of outdoor localization of a user as a regression problem. This formulation overcomes the limitations of other neural network based classification methods which estimates the position using a grid cell of pre-specified dimensions. We regress on the position directly which leads to better scalability when the testbed area is increased. Moreover, we introduce the usage of convolutional neural networks instead of fully connected neural networks to add more robustness to small changes in the environment. We evaluate our system on two different datasets to emphasize on the scalability of our regression approach. The testbeds are of size 0.147 km2 and 1.469 km2 . Our system achieves median localization error of 2.06m and 2.82m on each dataset respectively, outperforming current state-of-the-art outdoor cellular based systems by at least 877% improvement in the median localization error.
Environmental Sensitivity Evaluation of Neural Networks in Unmanned Vehicle Perception Module
Yuru Li (Peking University, China); Dongliang Duan (University of Wyoming, USA); Chen Chen and Xiang Cheng (Peking University, China); Liuqing Yang (Colorado State University, USA)
0
For autonomous driving of unmanned vehicles in intelligent transportation systems, multi-vehicle cooperative perception supported by vehicular networks can greatly improve the accuracy and reliability of the perception decisions. Currently, the perception decisions for a single vehicle are mostly provided by neural networks. Therefore, in order to fuse the perception decisions from multiple vehicles, the credibility of the neural network outputs needs to be studied. Among various factors, the environment is one of the most important affecting vehicles' perception decisions. In this paper, we propose a new evaluation criteria for the neural networks used in the perception module of unmanned vehicles. This criterion is termed as Environmental Sensitivity (ES), indicates the sensitivity of the network to environmental changes. We design an algorithm to quantitatively measure the ES value of different perception networks based on the extracted features. Experimental results show that our algorithm can well capture the sensitivity of the network in different environments and the ES values will be helpful to the subsequent decision fusion process.
Task Allocation for Mobile Crowdsensing with Deep Reinforcement Learning
Xi Tao and Wei Song (University of New Brunswick, Canada)
0
Mobile crowdsensing (MCS) is a new and promising paradigm of data collection in large-scale sensing and computing. A large group of users with mobile devices are recruited in a specific area to accomplish sensing tasks. An essential aspect of an MCS application is task allocation, which aims to efficiently assign sensing tasks to the recruited workers. Due to various resource and quality constraints, the MCS task allocation problem is often an NP-hard optimization problem. Traditional greedy or heuristic approaches are usually subject to performance loss in a certain degree so as to maintain tractability or accommodate special requirements such as incentive constraints. In this paper, we attempt to employ a deep reinforcement learning method to search for a more efficient task allocation solution. Specifically, we use a double deep Q-network (DDQN) to solve the task allocation problem as a path-planning problem with time windows. Our formulated problem takes into account location-dependency and time-sensitivity of sensing tasks, as well as the resource limits of workers in terms of maximum travelling distances. Simulations are conducted to compare the DDQN-based solution with two standard baseline solutions. The results show that our proposed solution outperforms the baseline solutions in terms of the platform's profit and the coverage of tasks.
Edge Caching Replacement Optimization for D2D Wireless Networks via Weighted Distributed DQN
Ruibin Li, Yiwei Zhao, Chenyang Wang and Xiaofei Wang (Tianjin University, China); Victor C.M. Leung (University of British Columbia, Canada); Xiuhua Li (Chongqing University, China); Tarik Taleb (Aalto University, Finland)
0
Duplicated download has been a big problem that affects the users' quality of service/experience (QoS/QoE) of current mobile networks. Edge caching and Device-to-Device communication are two promising technologies to release the pressure of repeated traffic downloading from the cloud. There are many researches about the edge caching policy. However, these researches have some limitations in the real scenarios. Traditional methods are lacking the self-adaptive ability in the dynamic environment and privacy issues will occur in centralized learning methods. In this paper, based on the virtue of Deep Q-Network (DQN), we propose a weighted distributed DQN model (WDDQN) to solve the cache replacement problem. Our model enables collaboratively to learn a shared predictive model. Trace-driven simulation results show that our proposed model outperforms some classical and state-of-the-art schemes.
Session Chair
Sung Whan Yoon (UNIST, Korea)
Session T4-S8
Cellular Networks and 5G
Conference
4:00 PM
—
5:30 PM KST
Local
May 27 Wed, 2:00 AM
—
3:30 AM CDT
Uplink Joint Detection: From theory to practice
Mohamed Amine Dridi (Nokia Bell Labs, France); Dora Boviz (Nokia, France); Eric Renault (Institut Mines-Telecom -- Telecom SudParis & Samovar UMR CNRS 5157, France); Laurent Roullet (Nokia Bell Labs, France); Ralf Klotsche (Nokia Bell Labs, Germany)
1
Ensuring a decent Quality of Experience (QoE) is fundamental for service providers, in particular mobile networks operators, when designing their current and future connectivity solutions. With this aim in view, they are compelled to cope with potential QoE detractors such as Inter-Cell Interference (ICI) which is expected to be a liability with a foreseen network densification. This issue was anticipated in Long Term Evolution (LTE) networks and many solutions leveraging cooperation schemes between the access nodes to alleviate the ICI's effects can be found in the literature, notably for uplink (UL) transmissions which pose a greater challenge. Among the cooperative models dealing with UL ICI, Joint Detection (JD) is particularly interesting since it promises substantial throughput gains while maintaining a high spectral efficiency. However, its practical feasibility is still unclear. In the following work, we propose a platform gathering a set of architectural, functional and technical requirements to endow realistic LTE networks with JD capabilities.
Nonlinear Digital Self-interference Cancellation with SVR for Full Duplex Communication
Mikail Yilan, Huseyin Ozkan and Ozgur Gurbuz (Sabanci University, Turkey)
2
Full duplex (FD) communication has attracted significant attention due to its potential for increasing the wireless link rate twofold without increasing the occupied bandwidth. For enabling FD communication, the self-interference (SI) signal at the transmitting radio should be suppressed down to the noise level. Despite SI cancellation applied at different stages via passive, analog and digital techniques, the current methods cannot sufficiently suppress SI at all power levels. Specifically at high power levels, the nonlinear behavior of the radio should be modeled within SI cancellation. In this paper, we propose a novel nonlinear digital cancellation approach by adapting support vector regression (SVR) for FD communication. The proposed SVR based nonlinear cancellation is integrated with linear cancellation and the digital SI cancellation algorithms are implemented and tested on a software defined radio set-up integrated with a monostatic antenna. With the proposed SVR based solution, up to 5 dB enhancement in total SI suppression is observed as compared to only linear digital cancellation, for the transmit power levels higher than 20 dBm. Moreover, for the same transmit power levels, up to 3 dB higher cancellation is achieved in comparison to the memory polynomial based nonlinear digital cancellation. Incorporating the proposed solution in the FD radio design only requires changes at the algorithmic level, which is implemented in software, hence there is no need for any hardware or circuitry modification. Additionally, the proposed nonlinear solution does not cause any extra communication overhead, since SVR models are to be learned only once for each transmit power level, then stored and re-used for the later transmissions.
A Real-Time Vendor-Neutral Programmable Scheduler Architecture for Cellular Networks
Wenhao Zhang, Zhouyou Gu, Wibowo Hardjawana and Branka Vucetic (The University of Sydney, Australia); Simon Lumb and David McKechnie (Telstra Corporation Ltd., Australia); Todd Essery (Telstra, Australia)
2
The current Downlink Shared Channel (DLSCH) resource scheduler for cellular networks has the following features: 1) it is integrated with an evolved NodeB (eNB) and 2) uses proprietary interfaces. The first causes a temporary outage whenever the scheduler logic is reprogrammed to accommodate traffic profiles that have different requirements, while the latter prevents multi-vendor interoperability. In this paper, we propose a real-time vendor-neutral programmable DLSCH scheduler architecture. The scheduler and eNB are separated into two binary files that communicate via an agent. The agent uses standard interfaces to interpret information from/to different eNB vendors in real time. The proposed architecture is implemented on two open source 3rd Generation Partnership Project standard- compliant eNB stacks from the OAI and SRS. Experimental results show that the proposed architecture addresses the real time and proprietary challenges mentioned above.
User Slicing Scheme with Functional Split Selection in 5G Cloud-RAN
Salma Matoussi (LIGM-CNRS, France); Ilhem Fajjari (Orange labs, France); Nadjib Aitsaadi (UVSQ Paris Saclay, France); Rami Langar (University Gustave Eiffel, France)
2
Next Generation 5G Radio Access Network (NG-RAN) is envisioned to integrate the slicing approach to build a flexible network supporting diverse use-cases with customized architectures, features and services. RAN processing functional splits have been standardized to add new deployment design capabilities and enhance cost efficiency. A further challenge consists in how to meet the multitude use-case's requirements while considering different design models in the physical infrastructure. Current related works are tackling the slice embedding problem from a cell-centric perspective. However, to achieve greater flexibility and better resource utilization, a user-centric approach should be more exploited. In this paper, we propose a SLICE-HPSO scheme that jointly harnesses radio, processing and link resources at the user level to build multiple user slices on top of the physical infrastructure. Our proposal is tailored to different user quality-of-service requirements and to the diverse functional splits resource requests. SLICE-HPSO is in compliance with the 3GPP and optimizes further the heterogeneous resource usage while meeting the scalability requirement.
Virtual Network Function Deployment Strategy in Clustered Multi-Mobile Edge Clouds
Yijing Liu, Gang Feng and Guanqun Zhao (University of Electronic Science and Technology of China, China); Zhuo Chen (Chongqing University of Technology, China); Shuang Qin (University of Electronic Science and Technology of China, China)
1
Software Defined Networking (SDN) and Network Function Virtualization (NFV) have been widely acknowledged as the fundamental architectural technologies for 5G and beyond networks by consolidating network functions into generalpurpose hardware. Meanwhile, emerging Mobile Edge Computing (MEC) technology provides a promising solution to fulfill service requirements on high reliability and low latency, by extending cloud computing to the edge of networks. To provision a specific type of service, a virtualized network topology, such as Service Function Chain or Service Function Graph is constructed by logically connecting a set of virtual network functions (VNFs). In SDN/NFV-based mobile networks with the MEC cluster, it is imperative to develop an effective VNF deployment strategy to support various services with diverse Quality of Service requirements. In this paper, we propose a VNF deployment strategy for clustered MEC, including VNF placement and routing schemes, with the aim to minimize the average delay of service flows. We first formulate the problem as a two- dimensional knapsack problem, which is NP-Hard. To provide an efficient solution, we develop an improved genetic simulated annealing algorithm. Numerical results show that the proposed strategy can significantly reduce the end-to-end service delay in comparison with the state-of-the-art solutions.
Session Chair
Jeongho Kwak (DGIST, Korea)