LEGGERO project focuses on design and prototype of next generation of far edge routers, i.e. small routers deployed at the network customers’ sites to support mission-critical applications.
LEGGERO improvements are on three directions:
- Lower cost and lower power consumption.
- Higher reliability with the help of artificial intelligence.
- Support for cloud applications
The management platform will be able to deploy distributed applications on a large fleet of nodes.
LEGGERO is part of Spoke 4 – Programmable Networks for Future Services and Media
Project PI: Giacomo Verticale
- LEGGERO developed tools to improve the knowledge and performance of linux routing by exploiting the eBPF technology. LEGGERO has tools for: measuring how much time a packet spends in the various networking functions of the linux kernel; providing high-speed communications between light VMs running in the same host; offloading to the kernel of the networking functions of cloud applications.
- LEGGERO developed a machine learning algorithm capable of predicting network latency and reliability in an SD-WAN deployment and choose the network connection that maximizes service reliability. The framework comprises the collection of latency and reliability metrics of each SD-WAN tunnel and then the configuration of the tunnel by using an SDN southbound protocol.
- LEGGERO is developing a management and orchestration platform for a fleet of edge routers and a central cloud domain capable of placing cloud-native workloads on the best node.
- LEGGERO is developing a prototype of the next generation router.
Papers
• A Zulfiqar et al. The Slow Path Needs an Accelerator Too! SIGCOMM Comput. Commun. Rev. 53, 1 (2023).https://doi.org/10.1145/3594255.3594259
• F. Parola et al. Comparing User Space and In-Kernel Packet Processing for Edge Data Centers. SIGCOMM Comput. Commun. Rev. 53, 1 (2023). https://doi.org/10.1145/3594255.3594257
• Farbod Shahinfar et al. Disaggregate Applications Along End-Host Data-Path. In Proceedings of the on CoNEXT Student Workshop 2023 (CoNEXT-SW '23). https://doi.org/10.1145/3630202.3630230
• Sebastiano Miano et al. Accelerating network analytics with an on-NIC streaming engine, Computer Networks, (2024) https://doi.org/10.1016/j.comnet.2024.110231
• S. Miano, et al., "Morpheus: A Run Time Compiler and Optimizer for Software Data Planes," in IEEE/ACM Transactions on Networking, doi: 10.1109/TNET.2023.3346286
Industrial results
• First prototype of the LEGGERO router manufactured by TIESSE.
• Open-source release of Zero Knowledge Middlebox https://github.com/bonsai-lab-polimi/zkIDS
• Open-source release of NETTO, an eBPF-based network diagnosis tool for Linux https://github.com/miolad/netto
• A Zulfiqar et al. The Slow Path Needs an Accelerator Too! SIGCOMM Comput. Commun. Rev. 53, 1 (2023).https://doi.org/10.1145/3594255.3594259
• F. Parola et al. Comparing User Space and In-Kernel Packet Processing for Edge Data Centers. SIGCOMM Comput. Commun. Rev. 53, 1 (2023). https://doi.org/10.1145/3594255.3594257
• Farbod Shahinfar et al. Disaggregate Applications Along End-Host Data-Path. In Proceedings of the on CoNEXT Student Workshop 2023 (CoNEXT-SW '23). https://doi.org/10.1145/3630202.3630230
• Sebastiano Miano et al. Accelerating network analytics with an on-NIC streaming engine, Computer Networks, (2024) https://doi.org/10.1016/j.comnet.2024.110231
• S. Miano, et al., "Morpheus: A Run Time Compiler and Optimizer for Software Data Planes," in IEEE/ACM Transactions on Networking, doi: 10.1109/TNET.2023.3346286
Industrial results
• First prototype of the LEGGERO router manufactured by TIESSE.
• Open-source release of Zero Knowledge Middlebox https://github.com/bonsai-lab-polimi/zkIDS
• Open-source release of NETTO, an eBPF-based network diagnosis tool for Linux https://github.com/miolad/netto
LEGGERO industrial partner is TIESSE, who has been involved in all the phases of the project innovation starting with the identification of the knowledge gap, the definition of the use cases and of the requirements, up to the validation of the technology. More specifically, NETTO will help in the identification of bottlenecks in the handling of packets in the prototype, while the Machine Learning component will improve the network reliability and will simplify the management of large fleets of devices.
LEGGERO has now identified three early adopters that will be involved in further validation in three specific use cases: paytech, internet access for business-critical applications, and smart grids. These companies will enjoy the benefits of LEGGERO innovations and will provide feedback to further improve the tools.
LEGGERO has now identified three early adopters that will be involved in further validation in three specific use cases: paytech, internet access for business-critical applications, and smart grids. These companies will enjoy the benefits of LEGGERO innovations and will provide feedback to further improve the tools.
Programs partners:
Cascade calls partners:
LEGGERO releases two open source tools:
1. The Zero Knowledge Middlebox https://github.com/bonsai-lab-polimi/zkIDS
2. NETTO, an eBPF-based network diagnosis tool for Linux https://github.com/miolad/netto
1. The Zero Knowledge Middlebox https://github.com/bonsai-lab-polimi/zkIDS
2. NETTO, an eBPF-based network diagnosis tool for Linux https://github.com/miolad/netto
1. Scientific publications: 5 / 10; RL = 1.2
2. Scientific publications with international scholars: 3 / 4; RL = 1.8
3. Webinars, workshops, or MOOCs: 0 / 3; RL = 0
4. Hackathons, theses, or student projects: 4 / 10; RL = 1
2. Scientific publications with international scholars: 3 / 4; RL = 1.8
3. Webinars, workshops, or MOOCs: 0 / 3; RL = 0
4. Hackathons, theses, or student projects: 4 / 10; RL = 1
1. Releases of Algorithms and Tools. 1 / 3; RL = 0.5
Researchers involved: 168
Collaboration proposals:
LEGGERO is looking for:
- early adopters of the next generation router,
- international scientific partners for participating in funded calls,
- institutions interested in hosting a hackathon on LEGGERO technology.
For any proposal of collaboration within the project please contact the project PI.
LEGGERO News: