PESCO aims to develop foundations and concrete technological enablers:
- for the future pervasive Internet, a complex environment, integrating communication, IoT and networking functionality
- for heterogeneous sensing integrated with communication functions, thanks to the abundance of wireless technologies with pervasive sensing capabilities
- based on a framework architecture and consistent performance evaluation for a pervasive Internet
- for making results sustainable, via linking its results with the international research and industrial communities
PESCO is part of Spoke 1 – Pervasive and Photonic Network Technologies and Infrastructures
Project PI: Andrea Passarella
Overall, the key results achieved until September 2024 are:
- Activities on all the scientific dimensions of the project. In particular, PESCO has achieved
- (i) novel solutions for supporting edge applications in heterogeneous pervasive Internet environments;
- (ii) solutions for efficient realisation of the FaaS concept at the edge of the pervasive Internet;
- (iii) an extended set of solutions for communication-efficient and resource-efficient AI in complex networked environments, with the inclusion of beyond-edge devices in the service provision fabric;
- (iv) initial results on novel forms of localisation and sensing also based on coordination among UAV swarms;
- (v) novel solutions of pervasive sensing integrated with diverse wireless communication technologies.
- definition of an initial version of the reference architecture and 10 shared reference use cases;
- Joint activities on 6 emerging topics among the project partners focusing on edge computing, decentralised AI, novel sensing and communication approaches
- Publication of 49 papers on the topics of PESCO, one of which received the Best Paper Award at an IEEE International Conference
- Contribution to 5 RESTART Grand Challenges
- Definition of 2 innovation use cases, in the areas of (i) novel architectures supporting edge computing through telco PoPs; and (ii) novel solutions for multiband and multidevice integrated communication and sensing.
[M1-M9]
Among the results mentioned in the section “Main Activities”, PESCO has defined the project use cases.
The 10 use cases are:
PESCO is significantly advancing the state of the art on all the scientific dimensions highlighted. Across the end of 2023 and the beginning of 2024, special focus has been devoted to launching transversal activities among the partners, which have also resulted in proposals for the RESTART Grand Challenges.
Specifically, the main areas where joint activities are progressing are the following:
[M21]
PESCO is significantly advancing the state of the art on all the scientific dimensions highlighted. Up until M21, special focus has been devoted to launching transversal activities among the partners, which have also resulted in proposals for the RESTART Grand Challenges. Specifically, the main areas where joint activities are progressing are the following:
Also thanks to these activities, PESCO is contributed four proposals for the following RESTART Grand Challenges
More specifically, the key results achieved so far can be summarised as follows:
Papers:
B. Picano, E. Vicario and R. Fantacci, "An Efficient Flows Dispatching Scheme for Tardiness Minimization of Data-Intensive Applications in Heterogeneous Systems," in IEEE Transactions on Network Science and Engineering, vol. 10, no. 6, pp. 3232-3241, Nov.-Dec. 2023, doi: 10.1109/TNSE.2023.3253501
Carlo Puliafito, Claudio Cicconetti, Marco Conti, Enzo Mingozzi, Andrea Passarella, Balancing local vs. remote state allocation for micro-services in the cloud–edge continuum, Pervasive and Mobile Computing, Volume 93, 2023, https://doi.org/10.1016/j.pmcj.2023.101808.
R. Mazzieri, J. Pegoraro and M. Rossi, "Attention-Refined Unrolling for Sparse Sequential micro-Doppler Reconstruction," in IEEE Journal of Selected Topics in Signal Processing, May 2024, doi: 10.1109/JSTSP.2024.3408041
Among the results mentioned in the section “Main Activities”, PESCO has defined the project use cases.
The 10 use cases are:
- Datacenter on edge;
- Drone swarm sensors synchronization;
- Future pervasive Internet for the Metaverse;
- Autonomous Intelligent Network of inertial sensors for real-time, high-precision environmental sensing (AISENSE-NET);
- Laser synchronization and secure key distillation in a free-space infrastructure;
- Smart Fleet Management for Logistic Systems;
- Computation and Communication Resources Optimization in Distributed Sensing and Learning;
- Indoor Human Movement Monitoring for Holographic Telepresence;
- Harbour surveillance: early identification and tracking of illegal activities;
- PESCO for Automated Mobile Robots in Smart Manufacturing Environments.
PESCO is significantly advancing the state of the art on all the scientific dimensions highlighted. Across the end of 2023 and the beginning of 2024, special focus has been devoted to launching transversal activities among the partners, which have also resulted in proposals for the RESTART Grand Challenges.
Specifically, the main areas where joint activities are progressing are the following:
- optimisation of the emerging Function as a Service (FaaS) paradigm for edge computing
- decentralised unlearning in pervasive networks
- the effects of network topologies and energy optimisation aspects of decentralised learning
- collaborative federated learning
- data processing and analysis of coherent distributed MIMO radar systems
- development of a near-field sensing testbed to validate near-field localization and sensing algorithms.
- Make artificial intelligence distributed and networked
- Sensing and protecting the world with communication signals
- Make your surroundings a programmable networked platform for customized services
[M21]
PESCO is significantly advancing the state of the art on all the scientific dimensions highlighted. Up until M21, special focus has been devoted to launching transversal activities among the partners, which have also resulted in proposals for the RESTART Grand Challenges. Specifically, the main areas where joint activities are progressing are the following:
- optimisation of the emerging Function as a Service (FaaS) paradigm for edge computing
- decentralised unlearning in pervasive networks
- the effects of network topologies and energy optimisation aspects of decentralised learning
- collaborative federated learning
- data processing and analysis of coherent distributed MIMO radar systems
- development of a near-field sensing testbed to validate near-field localization and sensing algorithms.
Also thanks to these activities, PESCO is contributed four proposals for the following RESTART Grand Challenges
- Challenge 7: Digitalize the environment for a sustainable world
- Challenge 10: Make the network a platform for programming and running applications
- Challenge 11: Sensing and protecting the world with communication signals
- Challenge 12: Make artificial intelligence distributed and networked
- Challenge 16: Connect and support everything and everyone that moves
More specifically, the key results achieved so far can be summarised as follows:
- Algorithms and protocols for the evolution of the pervasive Internet at the edge and beyond. A significant emphasis has been devoted to advanced edge systems, including integration of Digital Twin solutions also considering the key approach of Age-of-Information-based communications. Edge computing systems are becoming more and more a prime component of pervasive Internet system. PESCO is providing both novel algorithmic solutions as well as advanced practical platforms. This can have a significant societal/economic impact as the platforms developed in the project are ready to host the innovative algorithms which are developed, for prompt translation into pre-commercial products.
- Design and assessment of pervasive intelligence solutions. PESCO is working to assess the performance of decentralised pervasive learning in complex networked environments, also considering resource constraints of edge devices, and the need to unlearn (in addition to learn) over time observed patterns. This can have significant impact as the PESCO solutions match the significant academic and industrial trend on the decentralisation of intelligence and the need to design pervasive, resource-efficient distributed learning solutions.
- Advanced detection via cooperative UAV systems. PESCO has analyzed the imaging capability of a contactless radar system (e.g., mounted on drones) as a function of the collected amount of data. This is done by taking advantage of the concept of Interface Reflection Point (IRP) and the overall analysis allowed to determine a suitable sampling rate of the radar system in order to get satisfactory imaging results. Solutions based on MIMO radar have also been investigated. The potential impact of these results are important, as such solutions would allow more granular yet precise radar-based sensing mounted on specific devices (drones) that are one of the most important class of devices to make pervasive radar-based sensing a reality.
- New approaches for Pervasive sensing integrated with comms. PESCO has developed solutions for novel integrated and sensing technologies in typical edge computing environments. On the one end, new techniques for near-field sensing are developed with specific focus on vehicular environments. On the other end, solutions have been defined for edge computing assisted via communication-based sensing. The potential impact is high also in this case, as those techniques fall in one among the most innovative areas for post-5G networking.
Papers:
B. Picano, E. Vicario and R. Fantacci, "An Efficient Flows Dispatching Scheme for Tardiness Minimization of Data-Intensive Applications in Heterogeneous Systems," in IEEE Transactions on Network Science and Engineering, vol. 10, no. 6, pp. 3232-3241, Nov.-Dec. 2023, doi: 10.1109/TNSE.2023.3253501
Carlo Puliafito, Claudio Cicconetti, Marco Conti, Enzo Mingozzi, Andrea Passarella, Balancing local vs. remote state allocation for micro-services in the cloud–edge continuum, Pervasive and Mobile Computing, Volume 93, 2023, https://doi.org/10.1016/j.pmcj.2023.101808.
R. Mazzieri, J. Pegoraro and M. Rossi, "Attention-Refined Unrolling for Sparse Sequential micro-Doppler Reconstruction," in IEEE Journal of Selected Topics in Signal Processing, May 2024, doi: 10.1109/JSTSP.2024.3408041
In the period M1-M21 (until June 2024) PESCO included only one industrial partner, i.e., OpenFiber.
The main emphasis in the first half of the project has been on the focusing on the definition of the DataEdgers architecture. DataEdgers leverage the concept of single board computers and distributed computing, and will provide data and remote services in a decentralized way, on top of Open Fiber optical fiber network. The main goal of the new concept of datacenter is bringing data close to the edge, to grant data sovereignty at first and then to provide data localization and energy optimization. DataEdgers are meant to keep users’ data as close as possible to the end-users, reaching them thanks to the capillarity of POPs (point-of-presence) and PCNs (punto di consegna neutro), the most important facilities Open Fiber network is based on, allowing to implement remote services (e.g.: REST APIs, FAAS, etc.) with low latencies and pushing data towards edge computing technologies. Since DataEdgers are based on clusters of single-board-computers, the needed space allocation is going to be lower than the one required by traditional racks of common datacenters, and the energy consumptions are going to be lower as well. This also has impacts on operating costs, because DataEdgers do not require particular works on the buildings that are going to host them or require particular changes on already existing cooling systems. DataEdgers are going to be integrated within the OpenFiber optical fiber network infrastructure. A technological scouting to build the DataEdgers prototypes has been carried on and completed, to identify the hardware and software solutions that will be put on field to eventually build a test environment in 2 PoPs (point-of-presence) of the Open Fiber optical fiber network.
We have identified many receivers of our innovation: since we are developing a network of micro-datacenters based on clusters of single-board computers, to move data from cloud to the edge, everyone is using cloud services today could be interested in using our technologies tomorrow, trying to bring the end-user data as close as possible to the users themselves. This means e.g.: developers, open source software communities, consultancy companies, industries. Since DataEdgers are going to be available on Open Fiber infrastructure throughout the Italian territory, they are going to be provided in a very capillary way and this could bring even users that today are out of the cloud services because of the digital divide, to be interested to have digital twins or any other kind of digital services implemented on Open Fiber DataEdgers.
The main emphasis in the first half of the project has been on the focusing on the definition of the DataEdgers architecture. DataEdgers leverage the concept of single board computers and distributed computing, and will provide data and remote services in a decentralized way, on top of Open Fiber optical fiber network. The main goal of the new concept of datacenter is bringing data close to the edge, to grant data sovereignty at first and then to provide data localization and energy optimization. DataEdgers are meant to keep users’ data as close as possible to the end-users, reaching them thanks to the capillarity of POPs (point-of-presence) and PCNs (punto di consegna neutro), the most important facilities Open Fiber network is based on, allowing to implement remote services (e.g.: REST APIs, FAAS, etc.) with low latencies and pushing data towards edge computing technologies. Since DataEdgers are based on clusters of single-board-computers, the needed space allocation is going to be lower than the one required by traditional racks of common datacenters, and the energy consumptions are going to be lower as well. This also has impacts on operating costs, because DataEdgers do not require particular works on the buildings that are going to host them or require particular changes on already existing cooling systems. DataEdgers are going to be integrated within the OpenFiber optical fiber network infrastructure. A technological scouting to build the DataEdgers prototypes has been carried on and completed, to identify the hardware and software solutions that will be put on field to eventually build a test environment in 2 PoPs (point-of-presence) of the Open Fiber optical fiber network.
We have identified many receivers of our innovation: since we are developing a network of micro-datacenters based on clusters of single-board computers, to move data from cloud to the edge, everyone is using cloud services today could be interested in using our technologies tomorrow, trying to bring the end-user data as close as possible to the users themselves. This means e.g.: developers, open source software communities, consultancy companies, industries. Since DataEdgers are going to be available on Open Fiber infrastructure throughout the Italian territory, they are going to be provided in a very capillary way and this could bring even users that today are out of the cloud services because of the digital divide, to be interested to have digital twins or any other kind of digital services implemented on Open Fiber DataEdgers.
PESCO is addressing the vision of the next-generation Pervasive Internet. This is characterised by (i) pervasiveness of mobile devices and omnipresent connectivity; (ii) pervasiveness of data to be transmitted; (iiI) pervasiveness of intelligence to be extracted from data. To address these challenges, PESCO focuses on all the dimensions where communication of information and intelligence extracted from data join forces.
Publications
- Expected: >=1 journal publication/partner/year
- Accomplished: 26 journal publications + 31 conference publications
- Readiness: <100%
- Expected: >=15% joint publications/year within M18, >=30% within M36
- Accomplished: 8 joint publications over 57 (14%)
- Readiness: 47%
- Expected: 2 talks/partner/year reporting PESCO activities
- Accomplished: 25 talks delivered or already scheduled
- Readiness: 75%
- Expected: 0 (2 PoCs expected by the end of the project)
- Accomplished: 0 (Guiding use cases defined)
- Readiness: N.A. (work according to plan)
- Expected: 21 [1 plenary meeting/month]
- Accomplished: 41
- Readiness: <100%
MS1: Use cases & requirements, initial architecture draft
- Expected: 1 within M6
- Accomplished: 1 within M6
- Readiness: 100%
- Expected: 1 within M12
- Accomplished: 1 within M12
- Readiness: 100%
Researchers involved: 18 Full Time Equivalent
Collaboration proposals:
PESCO has defined a large set of technical Use Cases, as examples of the combination of technical components it will deliver, with respect to relevant application areas foreseen by project partners.
The project is happy to consider external collaborations with industries (and with SMEs in particular) to further enhance the current definition of the Use Cases, their extension, and/or possible trials based on the technical components planned for delivery.
The main areas for collaboration are:
- Novel paradigms, solutions and protocols for pervasive Internet, based on post-5G architectures
- Efficient pervasive AI at the edge of the network
- Novel IoT, sensing and localisation technologies
- Novel pervasive sensing solutions, also integrated with communication technologies
For any proposal of collaboration within the project please contact the project PI.
PESCO News: