Liquid Edge
Abstract
The goal of LIQUID_EDGE is to provide delay-sensitive cloud computing services to mobile devices through
(a) an efficient dynamic orchestration of computation, communication and caching (C3) resources;
(b) the use of innovative information-centric networking architectures facilitating mobility management; and (c) the exploitation of high data rate millimeter wave (mmWave) links.
The main idea is to crumble computation, communication, and networking resources down to a level such that they can move fast enough around the user to offer seamless service continuity, like a liquid pervasive computer. This is achieved through the dense deployment of mmWave radio access points equipped with edge computing capabilities and an intelligent orchestration of stateless micro-services and unikernel virtualization.
The project leverages on the synergic and dynamic interaction of the following enablers:
(i) MmWave advanced air interfaces, including multi-link and multi-RAT access, in a cell-free architecture;
(ii) Stateless microservice programming able to crumble a cloud service into small loosely coupled stateless functions that can be quickly deployed, booted and migrated thanks to their small footprint;
(iii) joint learning-based proactive and dynamic optimization of C3 resources
Project vision and objectives
Futuristic intelligent applications on mobile handsets are expected to handle a huge amount of data, extract relevant information and deliver the result with very low latency. Automatic Driving (AD), Virtual Reality (VR), Augmented Reality, multiple-cameras high-definition video capture and processing, are just some sample applications where there is the need to transfer large amount of data, process them,
and possibly take decisions with ultra high reliability and very low latency.
The main idea of LIQUID_EDGE is to face these challenging tasks is to crumble computation, communication, and networking resources down to a level where they can move fast enough around the user to offer seamless service continuity, like a liquid pervasive computer. This crumbling occurs concurrently at the communication and networking level, through the dense deployment of mmWave radio access
points (APs) equipped with edge computing capabilities, and at the computation level, through an intelligent orchestration of stateless microservices and unikernel virtualization.
The scenario envisioned in LIQUID_EDGE is consistent with the 5G roadmap and has the ambition to have a significant impact on the mobility management of intelligent applications requiring large data rate transfers, ultra reliability and very low latency. The methodology is strongly multi-layer and multi-disciplinary. It is multi-layer because the respect of low latencies entails a strong coupling between all layers, from the application all the way down to the physical layer. A revision of the entire architecture, including the split between control-plane and data-plane, and the native support to multi-RAT and multi-site connections is also necessary to overcome the instabilities of millimeter wave links.
The proposed methodology is also strongly multi-disciplinary as it relies on the convergence of expertise on communications, networking, virtualization and machine learning. In particular, guaranteeing strict latency constraints requires the development of online learning mechanisms fast enough to enable proactive resource allocation strategies.
One of the biggest challenges of 5G networks is to design a single platform able to support a variety of challenging vertical applications characterized by different requirements. This challenge is tackled by exploiting network resources in an intelligent way, including an efficient use of radio spectrum, multiple antennas, multi-Radio Access Technology (RAT) and multilink communications, guided by a
function-oriented design of the network functions enabling their virtualization and customizable deployment (slicing). Additionally, in order to reduce the end-to-end latency and backhaul traffic, cloud services need to be moved to the edge of the network, through the deployment of a dense set of in-network edge data centers. This architecture is called multi-access edge computing (MEC) [MEC] by ETSI, or mobile edge computing, and the edge data centers are referred to as Mobile Edge Hosts (MEH).
Despite the recent huge advances in 5G research, there are still fundamental issues concerning the efficient exploitation of mmWaves, whose instability poses many system level challenges, especially in terms of service continuity while users move from a cell to another or temporarily loose connection. LIQUID_EDGE plans to design and test an architecture that enables mobile users to get an efficient mmWave access to high data-rate cloud services experiencing very low end-to-end delay and mobile
service continuity. Our basic idea is to leverage on the synergic and dynamic interaction of the following enablers:
(i) MmWave advanced air interface providing gigabit radio access;
(ii) Multi-link and multi-RAT access to guarantee radio access continuity;
(iii) Cell-free networking architectures to ensure link stability and uniform performance;
(iv) Stateless microservice programming model for which a cloud service is crumbled into small loosely coupled stateless functions (e.g. Amazon AWS lambda);
(v) Unikernel virtualization technology for packaging the microservice as to make it autonomously runnable with small footprint and millisecond boot times [VEN16, MAR14];
(vi) A Liquid multi-access edge computing platform where microservices can be quickly deployed, booted and migrated thanks to their small footprint;
(vii) A data plane for accessing and transferring MEC microservices and related data based on Information Centric Networking (ICN) that provides name-based location-independent addressing, anycast, data
centric security and in-network caching;
(viii) A dynamic learning-based resource orchestration for the dynamic deployment of the microservices, in order to cope with communication, computation and caching (C^3) resources as aspects of a single
mobile architecture, following a user/application-centric perspective.
List of publications
G. Baruffa, A. Detti, L. Rugini, F. Crocetti, P. Banelli, G. Costante, P. Valigi, “AI-driven Ground Robots: Mobile Edge Computing and mmWave Communications at Work,” submitted to IEEE Open Journal of the Communications Society, accepted with minor, April 2024
G. Baruffa, L. Rugini, “Improved Channel Estimation and Equalization for Single-Carrier IEEE 802.11ad Receivers,” Radioengineering, vol. 32, no. 3, pp. 438–450, Sep. 2023.
G. Baruffa, L. Rugini, F. Frescura, P. Banelli, “Low-Complexity PAPR Reduction by Coded Data Insertion on DVB-T2 Reserved Carriers,” IEEE Access, vol. 11, pp. 73377–73393, July 2023.
G. Baruffa and L. Rugini, “Performance of LoRa-based Schemes and Quadrature Chirp Index Modulation,” IEEE Internet of Things Journal, vol. 9, no. 10, pp. 7759—7772, May 2022.
F. Binucci, P. Banelli, P. Di Lorenzo, S. Barbarossa. Adaptive resource optimization for edge inference with goal-oriented communications. EURASIP J. Adv. Signal Process. 2022, 123 (2022).
L. Funari, L. Petrucci and A. Detti, "Storage-Saving Scheduling Policies for Clusters Running Containers," in IEEE Transactions on Cloud Computing, vol. 11, no. 1, pp. 595-607, 1 Jan.-March 2023, doi: 10.1109/TCC.2021.3104662.
F. Binucci, P. Banelli, P. Di Lorenzo and S. Barbarossa, "Dynamic Resource Allocation for Multi-User Goal-oriented Communications at the Wireless Edge," 2022 30th European Signal Processing Conference (EUSIPCO), Belgrade, Serbia, 2022, pp. 697-701
F. Binucci, P. Banelli, P. Di Lorenzo, S. Barbarossa, "Multi-User Goal-Oriented Communications with Energy-Efficient Edge Resource Management," in IEEE Transactions on Green Communications and Networking, vol. 7, no. 4, pp. 1709-1724, Dec. 2023
F. Binucci and P. Banelli, "BER-Aware Dynamic Resource Management for Edge-Assisted Goal-Oriented Communications," ICASSP 2023 - 2023 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP), Rhodes Island, Greece, 2023, pp. 1-5
F. Binucci and P. Banelli, "Goal-Oriented Water-Filling for Dynamic Management of Edge-Assisted OFDM Communications," ICC 2023 - IEEE International Conference on Communications, Rome, Italy, 2023, pp. 5761-5766
F. Binucci, P. Banelli, P. Di Lorenzo and S. Barbarossa, "Analog versus Digital Pulse Amplitude Modulation for Goal-Oriented Wireless Communications," 2023 31st European Signal Processing Conference (EUSIPCO), Helsinki, Finland, 2023, pp. 1415-1419
G. Interdonato and S. Buzzi, "Joint Optimization of Uplink Power and Computational Resources in Mobile Edge Computing-Enabled Cell-Free Massive MIMO," in IEEE Transactions on Communications, vol. 72, no. 3, pp. 1804-1820, March 2024, doi: 10.1109/TCOMM.2023.3336355.
S. Buzzi et al., "LEO Satellite Diversity in 6G Non-Terrestrial Networks: OFDM vs. OTFS," in IEEE Communications Letters, vol. 27, no. 11, pp. 3013-3017, Nov. 2023, doi: 10.1109/LCOMM.2023.3320793
C. D'Andrea, G. Interdonato and S. Buzzi, "User-centric Handover in mmWave Cell-Free Massive MIMO with User Mobility," 2021 29th European Signal Processing Conference (EUSIPCO), Dublin, Ireland, 2021, pp. 1-5, doi: 10.23919/EUSIPCO54536.2021.9616361
G. Interdonato and S. Buzzi, "The Promising Marriage of Mobile Edge Computing and Cell-Free Massive MIMO," ICC 2022 - IEEE International Conference on Communications, Seoul, Korea, Republic of, 2022, pp. 243-248, doi: 10.1109/ICC45855.2022.9839124
G. Interdonato and S. Buzzi, "Conjugate Beamforming with Fractional-Exponent Normalization and Scalable Power Control in Cell-Free Massive MIMO," 2021 IEEE 22nd International Workshop on Signal Processing Advances in Wireless Communications (SPAWC), Lucca, Italy, 2021, pp. 396-400, doi: 10.1109/SPAWC51858.2021.9593193
C. D’Andrea, A. Garcia-Rodriguez, G. Geraci, L. G. Giordano and S. Buzzi, "Analysis of UAV Communications in Cell-Free Massive MIMO Systems," in IEEE Open Journal of the Communications Society, vol. 1, pp. 133-147, 2020, doi: 10.1109/OJCOMS.2020.2964983.
G. Faraci, S. Rizzo, G. Schembra, “Green Edge Intelligence for Smart Management of a FANET in Disaster-Recovery Scenarios,” IEEE Transactions on Vehicular Technology, Vol. 72, No. 3, March 2023.
F. Busacca, C. Grasso, S. Palazzo, G. Schembra, “A Smart Road Side Unit in a Microeolic Box to Provide Edge Computing for Vehicular Applications”, IEEE Transactions on Green Communications and Networking, Vol. 7, No. 1, March 2023.
G. Colajanni, P. Daniele, L. Galluccio, C. Grasso, G. Schembra,
“Optimizing FANET Lifetime for 5G Softwarized Network Provisioning,” IEEE Transactions on Network and Service Management, Vol. 19, No. 4, December 2022.
C.Grasso, R. Raftopoulos, G. Schembra, “H-HOME: A learning framework of federated FANETs to provide edge computing to future delay-constrained IoT systems,” Computer Networks, Volume 219, December 2022.
C.Grasso, R. Raftopoulos, G. Schembra, “Smart Zero-Touch Management of UAV-Based Edge Network,” IEEE Transactions on Network and Service Management, Volume 19, No. 4, December 2022.
F. Busacca, L. Galluccio, S. Palazzo, “A marketplace model for drone-assisted edge computing in 5G scenarios”, Elsevier Computer Networks, Volume 219, December 2022.
G. Colajanni, P. Daniele, L. Galluccio, C. Grasso, G. Schembra, “Service Chain Placement Optimization in 5G FANET-Based Network Edge,” IEEE Communications Magazine, November 2022.
C.Grasso, R. Raftopoulos, G. Schembra, “Slicing a FANET for heterogeneous delay-constrained applications,” Computer Communications, Volume 195, November 2022.
F. Busacca, G. Faraci, C. Grasso, S. Palazzo, G. Schembra, “Designing a Multi-Layer Edge-Computing Platform for Energy-Efficient and Delay-Aware Offloading in Vehicular Networks,” Elsevier Computer Networks, Volume 198, 2021.
C. Grasso, K. Eswar K.N., P. Nagaradjane, M. Ramesh, G. Schembra, "Designing the Tactile Support Engine to assist time-critical applications at the edge of a 5G network," Computer Communications, Volume 166, Pages 226-233, January 2021.
G. Faraci, C. Grasso, G. Schembra, "Design of a 5G Network Slice Extension With MEC UAVs Managed With Reinforcement Learning," in IEEE Journal on Selected Areas in Communications, vol. 38, no. 10, pp. 2356-2371, Oct. 2020.
G. Faraci, C. Grasso, G. Schembra, "Fog in the Clouds: UAVs to Provide Edge Computing to IoT Devices," ACM Transactions on Internet Technology, Vol. 20, No. 3, August 2020.
F. Busacca, S. Palazzo, R. Raftopoulos, G. Schembra, “MAD-FELLOWS: a Multi-Armed banDit Framework for Energy-efficient, Low-Latency job Offloading in robotic netWorkS”, Proc. IEEE INFOCOM NetRobiCS Workshop, Vancouver, May 2024.
A. Scarvaglieri, S. Palazzo, F. Busacca, “A lightweight, fully-distributed AI framework for energy-efficient resource allocation in LoRa networks”, Proc. DML-ICC 2023, Taormina, Italy, December 2023.
G. M. Cappello, G. Colajanni, P. Daniele, L. Galluccio, C. Grasso, G. Schembra, L. Scrimali, “ ODEL: an On-Demand Edge-Learning framework exploiting Flying Ad hoc NETworks (FANETs)”, Proc. of ACM Mobihoc 2023, Washington, November 2023
G. M. Cappello, G. Colajanni, P. Daniele, L. Galluccio, C. Grasso, G. Schembra, L. Scrimali, “Using FANETs for 6G Cloud-Native Slice Provisioning: A Marketplace Approach”, Proc. of 2023 IEEE Conference on Network Function Virtualization and Software Defined Networks, NFV-SDN 2023
F. Busacca, S. Palazzo, R. Raftopoulos, G. Schembra, " MANTRA: an Edge-Computing Framework based on Multi-Armed Bandit for Latency- and Energy-aware Job Offloading in Vehicular Networks" Proc. of IEEE Netsoft 2023, Madrid, Spain, June 2023.
L. Galluccio , C. Grasso , G. Maier , R. Raftopoulos, M. Savi , G. Schembra , S. Troia, “Reinforcement Learning for Resource Planning in Drone-Based Softwarized Networks”, Proc. of MedComNet 2022.
C. Grasso, R. Raftopoulos, G. Schembra, "Deep Q-Learning for Horizontal Job Offloading in a Fleet of MEC UAVs in 5G Environments" Proc. of IEEE Netsoft 2021, online, 28 June–2 July 2021.
C. Grasso, G. Schembra, "Decision Making Optimization for Job Offloading in Vehicular Edge Computing Networks," 2020 AEIT International Conference of Electrical and Electronic Technologies for Automotive (AEIT AUTOMOTIVE), Torino, Italy, November 18-20, 2020, pp. 1-6.
F. Busacca, L. Galluccio, S. Palazzo, “Drone-assisted edge computing: a game-theoretical approach”, Proc. IEEE INFOCOM WiSARN Workshop, online, July 2020.
G. Faraci, S. A. Rizzo, G. Schembra, "Unmanned aerial vehicles and wind generation serving isolated areas," 2020 International Symposium on Power Electronics, Electrical Drives, Automation and Motion (SPEEDAM), Sorrento, Italy, June 24-26, 2020, pp. 138-143.
F. Busacca, C. Cirino, G. Faraci, C. Grasso, S. Palazzo, G. Schembra, "Multi-Layer Offloading at the Edge for Vehicular Networks," 2020 Mediterranean Communication and Computer Networking Conference (MedComNet), Arona, Italy, June 17-19, 2020, pp. 1-8.
E. Conti, A. Piemontese, G. Colavolpe, and A. Vannucci, “Expectation propagation for flat-fading channels,” IEEE Wireless Commun. Letters, vol. 12, no. 11, pp. 1846–1850, Nov. 2023, DOI: 10.1109/LWC.2023.3295952
S. K. Dehkordi, L. Gaudio, M. Kobayashi, G. Caire, and G. Colavolpe, “Beam-space MIMO radar for joint communication and sensing with OTFS modulation,” IEEE Trans. Wireless Commun., vol. 22, no. 10, pp. 6737–6749, Oct. 2023, DOI: 10.1109/TWC.2023.3245207
L. Gaudio, G. Colavolpe, and G. Caire, “OTFS vs. OFDM in the presence of sparsity: A fair comparison,” IEEE Trans. Wireless Commun., vol. 21, no. 6, pp. 4410–4423, Jun. 2022, DOI: 10.1109/TWC.2021.3129975