ThD4 – Edge Computing
Thursday, 20 June 2019, 16:00-17:30, Room 4
Session chair: Maria Cuevas (BT Group Chief Technology Office, United Kingdom (Great Britain))
I-Hsun Chuang and Sun Rong Chen (National Cheng Kung University, Taiwan); Hsiang-Jen Tsai (NCKU, Taiwan); Mong-Fong Horng (National Kaohsiung University of Science and Technology & National Kaohsiung University of Applied Sciences, Taiwan); Yau Hwang Kuo (National Cheng Kung University, Taiwan)
With more and more IoT devices are deployed, our daily life will produce large amount of data. If we transmit these big data to the cloud servers directly, it would result in network traffic congestion in the cloud. In order to solve this problem, this paper using the Edge Computing to process data locally which means that processing data closer to where data be created. With the edge computing, it can reduce the latency from network transmission and it also can do real-time analytics. Although the edge computing solved some problem, it brought another problem here, how to decide what data should process in edge network and what should send to the cloud server. This problem will greatly affect the performance of the system. Therefore, this paper defined this problem as Dynamic Multiple Resource Management (DMRM) problem and discusses the solution for the resource allocation in dynamic environment with multiple resources. For the DMRM problem, this paper proposed the Multi-Resource Binary Particle Swarm Optimization (MR-BPSO) and also do some experiment comparing with other resource allocation methods. According to the experiments, the MR-BPSO has excellent performance and suitable for dynamic environment.
Che-Wei Hsu, Yung-Lin Hsu and Hung-Yu Wei (National Taiwan University, Taiwan)
The ultra-reliable and low latency communication (URLLC) and massive machine type communication (mMTC) in 5G are envisioned to support intelligent automation in the Factories-of-the-Future (FoF) environment. Mobile-edge computing (MEC) is thought of as a promising system for realization. However, the computation work cannot be done without reliable transmission. In this work, rather than simply investigating task offloading problem, the reliability of radio transmission is jointly considered under heterogeneous industrial IoT networks. A 2-tier MEC-cloud framework is provided, wherein the IoT mobile devices (MDs) are able to partition tasks and offload them to the MEC and the cloud server through the reliable transmission. A two-step algorithm named opportunity-cost-based offloading algorithm (OCBOA) is proposed to jointly optimize the allocation of communication and computation resources for task offloading with the minimum energy consumption and offloading failure probability. The experiments show that our low-complexity algorithm outperforms the other heuristic algorithms on resource allocation while satisfying the QoS requirements of the MDs.
Angelo Tulumello (CNIT / University of Rome Tor Vergata, Italy); Giacomo Belocchi (CNIT/University of Rome Tor Vergata, Italy); Marco Bonola (University of Rome “Tor Vergata”, Italy); Salvatore Pontarelli (National Inter-University Consortium for Telecommunications (CNIT), Italy); Giuseppe Bianchi (University of Rome “Tor Vergata”, Italy)
Offloading to the edge a subset of the cloud services requested by users is a very appealing solution to reduce the bandwidth pressure, minimize latency and improve the overall quality of experience of mobile users. Achieving this goal is technically challenging, since the network functionalities needed to manage this offloading are not trivial, and present significant requirements in terms of scalability and speed. We propose to realize these network functionalities directly in the dataplane, exploiting the characteristic of FlowBlaze, a novel stateful programmable dataplane. Furthermore, we show how the network functions for the traffic offload can be easily expressed using an ad-hoc domain specific language developed for the description of per-flow stateful network functions.
Emeka Emmanuel (LSBU, United Kingdom (Great Britain)); Saptarshi Ghosh, Muddesar Iqbal and Tasos Dagiuklas (London South Bank University, United Kingdom (Great Britain)); S Mumtaz (GS-lda, Portugal); Anwer Al-Dulaimi (EXFO Inc., Canada)
Multi-Access Mobile Edge Computing (MEC) is proclaimed as a key technology for reducing service processing delays in 5G networks. Caching on MEC will decrease service latency and improve data access by allowing direct content delivery through the edge without fetching content from the remote server, Caching on MEC is also deemed as an effective approach that guarantees more reachability due to proximity to end-users. This paper proposes a novel hybrid content caching replacement algorithm in MEC to increase its caching efficiency where future request references are predicted using a polynomial fit algorithm along with Lagrange interpolation. Additionally, a distributed co-operative caching algorithm to improve data access within MECs. Experimental results have shown that the proposed scheme obtains more cache hits and lesser average CPU utilization due to its selective caching approach when compared with existing traditional cache replacement algorithms.
Marc Mollà Roselló (Ericsson Spain, Spain)
We present a practical approach of how we can use deep learning models to improve 5G network service. We demonstrate the potential of a deep Q-network agent for solving a traffic management problem, which can be applied for optimizing the network in multi-path scenarios. We use for the demonstration a multi-path QUIC implementation and we train an agent for improving the algorithm that selects the optimal path, with results in a better utilization of the network resources.