Partial Replication for Delay-Optimal Distributed Edge Computing

The ever-increasing scale and complexity of the computing tasks arising from various mobile applications has fostered wide research interests in distributed edge computing. Coded edge computing is among the recent advancements in this area, which can effectively mitigate the task processing delay caused by straggling edge nodes (ENs). In edge computing, as the mobile user usually tends to offload to closer ENs to save transmit power, the adversary may stealthily infer user location by exploiting this feature. Although there have been some pioneering works on location privacy-aware offloading, they mainly focused on the single EN scenarios and may not be directly applicable to coded edge computing that involves multiple ENs. To the best of our knowledge, the location privacy issue in coded edge computing still remains largely unexplored. With this consideration, a sequential hypothesis testing based location inference attack is identified in this work to reveal the potential vulnerability of existing coded edge computing methods. Besides, a countermeasure based on dynamic EN selection is proposed, together with a location privacy-aware coded offloading scheme based on the generic Lyapunov optimization framework. In addition to analysis, simulation results are provided to justify the effectiveness of the proposed scheme.

An edge computing based anomaly detection method in IoT industrial sustainability

In edge computing, edge nodes are hierarchically arranged, while data transmission is allowed among edge nodes. Because end devices are the closest to raw data sources, they usually submit requests with sensitive information to the edge nodes. Privacy issues occur during the transmission process. In addition, when an end user sends a request to edge nodes, the end user connects to the closest edge node initially, and the request is passed to the upper-layer edge nodes for further processing when the resources of the initial nodes are exhausted. The resource limitation includes computation, storage capabilities, and the number of users. For example, if a user generates one request to the connected edge node exceeding its computation power, the upper-layer edge node will be involved in the operation. Additionally, the information will be transmitted to the other edge nodes, which might be malicious. Thus, privacy preservation is necessary to prevent privacy leakage from data transition among multiple edge nodes in hierarchical structures.

This paper mainly focuses on the volume calculation of materials in the warehouse where sand and gravel materials are stored and monitored whether materials are lacking in real-time. Specifically, we proposed the sandpile model and the point cloud projection obtained from the LiDAR sensors to calculate the material volume. We use distributed edge computing modules to build a centralized system and transmit data remotely through a high-power wireless network, which solves sensor placement and data transmission in a complex warehouse environment. Our centralized system can also reduce worker participation in a harsh factorial environment. Furthermore, the point cloud data of the warehouse is colored to visualize the actual factorial environment. Our centralized system has been deployed in the real factorial environment and got a good performance.

Edge computing is a novel network architecture that is in proximity to the end devices in an Internet of Things (IoT). As the IoT becoming a major factor in our daily life, provisioning a low response time of the services to IoT users through edge computing is an important problem. Caching necessary program data for the task in an edge node effectively reduces the response time of the computation task. However, due to the increase of IoT users and devices, it is noteworthy that limited-resource edge nodes would receive a number of tasks, having a heavy burden of processing the requests. Therefore, the limited resource and caching space at cloudlet need the careful design of the caching algorithm to utilize the space of multiple edge nodes and relieve the burden of computations. In this paper, we propose a cooperative program caching system that makes different edge nodes cooperatively store program data and cache the replicas of the data requested frequently to handle a number of requests from IoT users. In particular, we develop a cooperative caching algorithm that caches the appropriate number of data replicas depending on the number of requests on each cloudlet and the popularity of the data to minimize the response time. The simulation results show that the proposed cooperative caching algorithm can effectively reduce the response time for IoT users compared to other existing algorithms.

The edge computing works as an interface between the real world data, sensor nodes and cloud devices and provides intended security and privacy features. Use of smart wearable devices embedded with modern chipsets and leveraged with edge computing provides near real time monitoring, display diagnostics of state of the patient health and prevents spread of disease, even if the doctor is distantly located. In this paper, a secure smart health care system is proposed that is supported by multiple physiological sensors, smart devices and edge nodes at patient and doctor end to collect, protect, process, analyze and leverages cloud facilities to manage health records of patients. The analyzed data and requisites data are made available to the patient's smart devices. The simulation results show that the proposed system demands low latency for patients under different medical condition to enable fast and real-time diagnosis from healthcare professionals.

Recent advances in machine learning and deep learning have enabled many existing applications in smart cities, autonomous systems, and wearable devices. These applications often demand scalable real-time cognitive intelligence and on-the-spot decision making. Current computer systems have been customized for a cloud computing paradigm which often does not meet latency constraints and scalability requirements. To address the limitations of the cloud computing paradigm, the general trend is toward shifting the computation next to data producers at the edge. However, the edge computing paradigm is in the very early stages. Many system-level aspects of edge computing, including algorithms mapping and partitioning across edge computing resources (edge server, and edge nodes) are unknown. New research is required to understand and quantify design dimensions for edge computing.This paper presents a novel edge computing infrastructure for distributed real-time video analytics. This paper presents a holistic solution for co-designing application and edge infrastructure, including edge nodes and edge servers, to enable scalable real-time Artificial Intelligence (AI)/Deep Learning (DL) video analytics across many cameras. For experimental results and evaluation, we focus on the case study of object re-identification across many cameras, which is composed of object detection/classification (TinyYOLOv3), feature extraction, local re-identification, and global re-identification kernels. We evaluate the edge system under three different task mapping and resource allocation configurations. The results present that with the edge nodes (video cameras) more than 32, the only scalable solution is to perform detection/classification (TinyYOLOv3), feature extraction, local re-identification on the edge nodes next to cameras, and execute global re-identification on edge server.

Coded computation has attracted great interests as a promising technique to cope with straggling computing nodes in mobile edge computing (MEC) networks. Contrary to the existing coded computation schemes developed with a fixed network topology, this paper studies a MDS coded computation for random networks. Specifically, we put forth maximum distance separable (MDS) coded task offloading and investigate its MDS coded computing gain by deriving the average successful retrieval probability with stochastic geometry in random wireless edge computing networks, where it encodes the original task into multiple equal and small sized MDS coded sub-tasks and offloads their subset to edge computing nodes for computation. We also identify a tradeoff between the latency in processing a sub-task at an edge node and the minimal number of edge nodes required to retrieve the original task output, according to the size of MDS coded sub-tasks. To efficiently control the tradeoff, we determine the desirable size of MDS coded sub-tasks in a semi-closed form to maximize the average successful retrieval probability for regime 1 and regime 2 networks, which correspond to the cases that communication latency is negligible compared to computation latency and that computation latency is negligible compared to communication latency, respectively, and develop an efficient algorithm with low search complexity for a general environment. Our numerical results reveal that the proposed scheme outperforms the other conventional task offloading schemes such as partial task offloading and replication task offloading in terms of average successful retrieval probability.

This thesis focuses on how to optimize the placement of the Edge Computing infrastructure for upcoming 5G networks. To this aim, the core contributions of this research are twofold: 1) a novel heuristic called Hybrid Simulated Annealing to tackle the NP-hard nature of the problem and, 2) a framework called EdgeON providing a practical tool for real-life deployment optimization. In more detail, Edge Computing has grown into a key solution to 5G latency, reliability and scalability requirements. By bringing computing, storage and networking resources to the edge of the network, delay-sensitive applications, location-aware systems and upcoming real-time services leverage the benefits of a reduced physical and logical path between the end-user and the data or service host. Nevertheless, the edge node placement problem raises critical concerns regarding deployment and operational expenditures (i.e., mainly due to the number of nodes to be deployed), current backhaul network capabilities and non-technical placement limitations. Common approaches to the placement of edge nodes are based on: Mobile Edge Computing (MEC), where the processing capabilities are deployed at the Radio Access Network nodes and Facility Location Problem variations, where a simplistic cost function is used to determine where to optimally place the infrastructure. However, these methods typically lack the flexibility to be used for edge node placement under the strict technical requirements identified for 5G networks. They fail to place resources at the network edge for 5G ultra-dense networking environments in a network-aware manner. This doctoral thesis focuses on rigorously defining the Edge Node Placement Problem (ENPP) for 5G use cases and proposes a novel framework called EdgeON aiming at reducing the overall expenses when deploying and operating an Edge Computing network, taking into account the usage and characteristics of the in-place backhaul network and the strict requirements of a 5G-EC ecosystem. The developed framework implements several placement and optimization strategies thoroughly assessing its suitability to solve the network-aware ENPP. The core of the framework is an in-house developed heuristic called Hybrid Simulated Annealing (HSA), seeking to address the high complexity of the ENPP while avoiding the non-convergent behavior of other traditional heuristics (i.e., when applied to similar problems). The findings of this work validate our approach to solve the network-aware ENPP, the effectiveness of the heuristic proposed and the overall applicability of EdgeON. Thorough performance evaluations were conducted on the core placement solutions implemented revealing the superiority of HSA when compared to widely used heuristics and common edge placement approaches (i.e., a MEC-based strategy). Furthermore, the practicality of EdgeON was tested through two main case studies placing services and virtual network functions over the previously optimally placed edge nodes. Overall, our proposal is an easy-to-use, effective and fully extensible tool that can be used by operators seeking to optimize the placement of computing, storage and networking infrastructure at the users’ vicinity. Therefore, our main contributions not only set strong foundations towards a cost-effective deployment and operation of an Edge Computing network, but directly impact the feasibility of upcoming 5G services/use cases and the extensive existing research regarding the placement of services and even network service chains at the edge.

Application-aware computation offloading in edge computing networks

Due to the urgent emergence of computation-intensive intelligent applications on end devices, edge computing has been put forward as an extension of cloud computing, to satisfy the low-latency requirements of these applications. To process heterogenous computation tasks on an edge node, the corresponding services should be placed in advance, including installing softwares and caching databases/libraries. Considering the limited storage space and computation resources on the edge node, services should be elaborately selected and deployed on the edge node and its computation resources should be carefully allocated to placed services, according to the arrivals of computation workloads. The joint service placement and computation resource allocation problem is particularly complicated, in terms of considering the stochastic arrivals of tasks, the additional latency incurred by service migration, and the waiting time of unprocessed tasks. Benefiting from deep reinforcement learning, we propose a novel approach based on parameterized deep Q networks to make the joint service placement and computation resource allocation decisions, with the objective of minimizing the total latency of tasks in a long term. Extensive simulations are conducted to evaluate the convergence and performance achieved by our proposed approach.

SummaryWith the advent of the Internet of things era, cloud computing platforms will face the challenges of massive equipment requirements for access, massive data, insufficient bandwidth, and high power consumption. Edge computing, as a new technology, makes it possible to stream media over the edge network. However, many problems related to file sharing among edge nodes and to the files provided by cloud servers exist. We propose a novel replica placement strategy for mobile media streaming in edge computing (RPME) to address the aforementioned problem. First, we introduce a multilevel replica placement model in the RPME. In the RPME, we acquire the user information and the user‐item rating matrix when the user requests the media data. At the server, we cluster the users according the user information, update the user‐item rating matrix, and then generate a replica recommendation sequence. For when the server submits the replica recommendation sequence to the edge node considering the constraints of the edge node storage capacity and the limitations on the service capacity of the requested replica, we propose an effective replica placement mechanism in the RPME. To show the benefits of the RPME, we also present several experiments to prove its validity.

Both network function virtualization (NFV) and edge computing (EC), especially the latter, are attracting more and more attention in recent years. A growing number of network service providers are migrating their services from the cloud to the edge for better QoS services, while the recent researches on NFV also concentrate on deploying NFV services in edge computing networks. However, NFV deployment in edge networks is a troublesome challenge and is fairly alien from conventional NFV deployment problems in data centres. Edge network differs from the data center network in the following two aspects: firstly, edge nodes are constrained in computing capacity, and secondly, the network connections between edge nodes are unstable and dynamic, which may show large variance over time. This means edge computing should be designed for high-efficient use of physical edge nodes’ resources. To address the challenges above, we investigate a new NFV Service Chain Placement problem in edge computing environments (NSCP-EC) in this paper. We first prove that the NSCP-EC problem is NP-complete. Then we propose a new metric which can better measure the capacity utilization rate of physical resources, and analyze its advantages with details. Based on the new metric, we propose two heuristic but efficient algorithms called MINI and MINI-tree. To confirm the performance of the two algorithms, we conduct simulations. The result demonstrates that MINI gains an advantage over genetic algorithm (GA) and MINI-tree orevails over MINI in tree topology conditions in the aspects of physical resource utilization, acceptance rate and running time. Both theoretical analysis and simulation results confirm the feasibility of the algorithms.

Multi-access/Mobile Edge Computing (MEC) and fog computing are promising techniques to satisfy low latency requirements for emerging next generation applications. Moving computation entities closer to a user could reduce the overall serving latency. In terms of green communications, given the latency constraint, how to minimize the power consumption at the user equipment (UE) and the edge node (EN) sides is important. Considering several edge nodes, partially offloading a user's task to one or more edge nodes is key. In this paper, a multi-node partial task offloading MEC scenario is discussed, in which UEs locally compute the task and share the remainder with other edge nodes. In addition, a green task distribution algorithm which minimizes the system power consumption is proposed, considering queueing, transmitting and computing delay. The simulation results show that the proposed algorithm minimizes the power consumption while meeting the latency requirements, and the power saving efficiency outperforms a binary offloading strategy. Moreover, the coupling effects between the latency requirements, offloading signal strength within the edge nodes, computation capability of edge nodes and the number of subcarriers used to transmit the offloading task are discussed.

Edge computing and machinery automation application for intelligent manufacturing equipment