Session Demo-3

Demo Session 3

8:00 PM — 10:00 PM EDT
May 12 Wed, 8:00 PM — 10:00 PM EDT

An Interactive and Immersive Remote Education Platform based on Commodity Devices

Jiangong Chen (University of Rhode Island, USA); Feng Qian (University of Minnesota, Twin Cities, USA); Bin Li (University of Rhode Island, USA)

Virtual reality (VR) holds a great potential to provide interactive and immersive learning experiences for students in remote education by using existing mobile devices, which is extremely meaningful during the current pandemic. In such a VR application, satisfactory user experience requires: 1) high-resolution panoramic image rendering; 2) high frame rate; 3) synchronization among users. This requires that either mobile devices perform fast image rendering or today's wireless network can support multi-Gbps traffic with extremely low delay, neither of which is the case in current practice. In this demo, we develop a platform for interactive and immersive remote education based on commodity devices, where a server performs rendering to ensure that the rendered images have high-resolution (\(2560\times 1440\) pixels) and are displayed at a high frame rate (60 frames per second) on the client-side. We further leverage motion prediction to overcome the diverse round-trip time (RTT) between a server and users and ensure synchronization among users (average 9.2 ms frame latency difference among users), which improves at least 60% and 20% compared to the existing local-rendering and server-rendering methods, respectively.

Smart Contract-enabled LightChain Test Network

Yahya Hassanzadeh-Nazarabadi (DapperLabs, Canada); Kedar Kshatriya (Savitribai Phule Pune University, India); Oznur Ozkasap (Koc University, Turkey)

LightChain is the first Distributed Hash Table (DHT)-based blockchain with a logarithmic asymptotic operational complexity, and a distributed storage layer, which preserves its integrity under the corrupted majority power of nodes. Running smart contract-based transactions, however, was a missing feature in the original implementation of LightChain. In this demo paper, we present the software architecture of our open-source smart contract-enabled test network for LightChain.

SwiftS: A Dependency-Aware and Resource Efficient Scheduling for High Throughput in Clouds

Jinwei Liu (Florida A&M University, USA); Long Cheng (North China Electric Power University, China)

An increasing number of large-scale data analytics frameworks moves towards larger degrees of parallelism aiming at high throughput guarantees. It is challenging to design a scheduler with high throughput and high resource utilization due to task dependency and job heterogeneity. The state-of-the-art schedulers in cloud/datacenters cannot well handle the scheduling of heterogeneous jobs with dependency constraints (e.g., dependency among tasks of a job) for simultaneously achieving high throughput and high resource utilization. We propose SwiftS, a dependency-aware and resource efficient scheduling for high throughput in clouds.

Multi-domain MEC orchestration platform for enhanced Back Situation Awareness

Nina Slamnik-Krijestorac (University of Antwerp, IDLab-imec, Belgium); Girma Mamuye Yilma, Faqir Zarrar Yousaf and Marco Liebsch (NEC Laboratories Europe GmbH, Germany); Johann M. Marquez-Barja (University of Antwerpen & imec, Belgium)

Network Function Virtualization (NFV) and Multi-Access Edge Computing (MEC) are among the key technology pillars of 5G systems and beyond for fostering and enhancing the performance of new and existing use cases. In the context of public safety, 5G offers great opportunities towards enhancing mission-critical services, by running network functions at the network edge to provide reliable and low-latency services. This demo introduces an on-demand Back Situation Awareness (BSA) application service, in a multi-domain scenario, enabling early notification for vehicles of an approaching Emergency Vehicle (EmV), indicating its Estimated Time of Arrival (ETA). The application provides the drivers ample time to create a safety corridor for the EmV to pass through unhindered in a safe manner thereby increasing the mission success. For this demo, we have developed an orchestrated MEC platform on which we have implemented the BSA service following modern cloud-native principles, based on Docker and Kubernetes.

FIND: an SDR-based Tool for Fine Indoor Localization

Evgeny Khorov (IITP RAS, Russia); Aleksey Kureev (IITP RAS & MIPT, Russia); Vladislav Vladimirovich Molodtsov (IITP RAS, Russia)

An indoor localization approach uses Wi-Fi Access Points (APs) to estimate the Direction of Arrival (DoA) of the Wi-Fi signals. This paper demonstrates FIND, the tool for Fine INDoor localization based on software-defined radio (SDR), which receives Wi-Fi frames in the 80 MHz band with four antennas. To the best of our knowledge, it is the first-ever prototype that extracts from such frames data in both frequency and time domains to calculate DoA of Wi-Fi signals in real-time. Apart from other prototypes, we retrieve from frames comprehensive information that could be used to DoA estimation: all preamble fields in the time domain, Channels State Information (CSI) and signal-to-noise ratio. Using our device, we collect a dataset for comparing different algorithms estimating angle of arrival in the same scenario. Furthermore, we propose a novel calibration method, eliminating the constant phase shift between receiving paths caused by hardware imperfections. All calibration data, as well as a gathered dataset with various DoA in an anechoic chamber and in a classroom, are provided to facilitate further research in this area.

SDR-based Testbed for Real-time CQI Prediction for URLLC

Kirill Glinskiy (Moscow Institute of Physics and Technology, Russia); Aleksey Kureev (IITP RAS & MIPT, Russia); Evgeny Khorov (IITP RAS, Russia)

Ultra-reliable Low-Latency Communication (URLLC) is a key feature of 5G systems. The quality of service (QoS) requirements imposed by URLLC are less than 10ms delay and less than 10-5 packet loss rate (PLR). To satisfy such strict requirements with minimal channel resource consumption, the devices need to accurately predict the channel quality and select Modulation and Coding Scheme (MCS) for URLLC in a proper way. This paper presents a novel real-time channel prediction system based on Software-Defined Radio that uses a neural network. The paper also describes and shares an open channel measurement dataset that can be used to compare various channel prediction approaches in different mobility scenarios in future research on URLLC.

Experimenting in a Global Multi-Domain Testbed

Esteban Municio (University of Antwerp - imec, Belgium); Mert Cevik (RENCI - UNC Chapel Hill, USA); Paul Ruth (UNC-CH, USA); Johann M. Marquez-Barja (University of Antwerpen & imec, Belgium)

New AI-based and ULLC applications are demanding novel network management approaches that are capable to cope with unprecedented levels of flexibility, scalability and energy efficiency. In order to make these use cases tangible, network management solutions aim to rely on multi-domain, multi-tier architectures that permit complex end-to-end orchestration of resources. However, current research on scheduling functions and task-offloading algorithms often focus on one single-domain, and the exploration of large-scale inter-operable solutions becomes a challenge. Fortunately for the researchers, a number of available testing facilities deployed at different geographical location in the world can be integrated to be used as a joint multi-domain infrastructure. In this demo, we present a hands-off experience of how to integrate different high-performance testbeds, located in USA, Belgium and The Netherlands, to enable multi-domain large-scale experimentation. We show end-to-end performance characteristics of the testbed integration and describe the main takeaways and lessons learned to drive researchers towards successful deployments in such end-to-end global infrastructure.

FLEX: Trading Edge Computing Resources for Federated Learning via Blockchain

Yang Deng (University of North Carolina, Charlotte, USA); Tao Han (University of North Carolina at Charlotte, USA); Ning Zhang (University of Windsor, Canada)

Federated learning (FL) algorithms provide privileges in personal data protection and information islands elimination for distributed machine learning. As an increasing number of edge devices connected in networks, we still see a lot of computing resources and data remaining underutilized and there is no platform for users to trade FL tasks. In this demonstration, we propose a blockchain-based federated learning application trading platform called FLEX, on which users can buy and sell machine learning models with no sacrifice of data privacy. We design FLEX in a highly distributed and scalable manner. We separated the data plane and control plane in the platform. In FLEX, trading mechanisms and FL algorithms are deployed in smart contracts of the blockchain. Control messages and trading information are well protected in the blockchain.With FLEX, we realize a distributed trading platform for executing FL tasks.

Session Chair

Linke Guo (Clemson University, United States)

Made with in Toronto · Privacy Policy · INFOCOM 2020 · © 2021 Duetone Corp.