EDGELESS

EDGELESS is pushing the boundaries of real-time edge computing! We are evaluating our solutions across three pilots: ?? Autonomous Smart City Surveillance – Enhancing citizen security with ultra-fast object detection & event inference. ?? Internet of Robotic Things (IoRT) – Enabling autonomous robots to adapt seamlessly in agile production. ?? HealthCare Assistant (HCA) – Using AI to detect anomalies & support personalized healthcare. ?? Stay tuned for updates! #EDGELESS #EdgeCloud

Check out this new feature added to #EDGELESS: transfer time measurement??. EDGELESS is a framework to manage and optimize the deployment of stateful #FaaS applications in edge-cloud computing infrastructures. For the optimization part, it is of paramount importance to have fine-grained and timely information on the performance of the functions deployed, which realize the user service. This is done by means of a telemetry sub-system ?? that gathers per-event (one event == a function call) samples, periodically sent by the edge/cloud nodes to the ε-ORC, which oversees the orchestration process locally. Up until v1.0.0, we only tracked the execution time, i.e., the time it takes a node to handle a function call from when processing starts?. Recently, we improved the richness of information collected by also measuring the transfer time, which is computed from when an event is generated (= the caller invokes the next function) to when the callee starts handling it. In addition to including any network latency associated with the inter-process communication between nodes, this metric also includes the time that an event might be spent before the callee can handle it, if there are pending events to be processed. Try this today from the main branch of our open-source repository at https://bit.ly/3YPSZaf #EDGELESS #FaaS #EdgeComputing #Cloud #Optimization

We are on Mastodon – Join Us! ?? Our community is growing, and we’d love for you to be part of it! Follow us on Mastodon to stay connected, explore our latest updates, and engage in insightful discussions. ?? Follow us here: https://bit.ly/4fNr13v #EDGELESS

Our paper “Energy-Efficient Task Computation at the Edge for Vehicular Services” has been accepted at IEEE/IFIP NOMS 2025! ?? In this work, we explore how Multi-Access Edge Computing (MEC) can support vehicular services by enabling efficient task offloading. We tackle the challenge of energy-efficient computation, considering real-world car mobility patterns from a leading European mobile network operator. ?? Key Highlights: ? Optimized task computation and offloading with multi-agent reinforcement learning ? Energy savings of 47% (static) and 14% (mobile) compared to state-of-the-art solutions ? Reduced task interruptions and user dissatisfaction Excited to present this at NOMS 2025 and discuss how our approach improves efficiency in autonomous driving and smart mobility! ??? ?? Authors: Paniz Parastar, Giuseppe Caso, Jesus Alberto Oma?a Iglesias, Andra Lutu, ?zgü Alay ?? Affiliations: University of Oslo, Karlstad University, Telefónica Research Looking forward to engaging discussions! Feel free to reach out. ?? #MEC #EdgeComputing #AI #ReinforcementLearning #VehicularServices #NOMS2025 #EnergyEfficiency

#Worldine, coordinator of the European #EDGELESS project, has implemented a CI pipeline for EDGELESS that builds #Dockerimages automatically when a release is generated for the main system components: Orchestrator, Controller and Node.This allows users to deploy an EDGELESS cluster easily executing just a few commands. More information here: https://bit.ly/3GEoBFE #EDGELESS #AI #serverlesscomputing

Check out the recent presentation of “Benchmarking an EDGELESS Cluster for Serverless Edge Computing Applications” at #ML4ECS. EDGELESS is a complex system, which has: ? a lot of workflows, running in ? multiple domains, each hosting ? many nodes, each running ? several function/resource instances. Therefore, it is not an easy task to evaluate the performance of an #EDGELESS system in controlled and repeatable conditions. For this we are developing edgeless_benchmark, which interacts as a client with the real system (applictions can be mixed with real services) and creates workflows in a repeatable manner to: - Study the impact of a new algorithm (e.g., orchestration function). - Provision the system. - Assess what-if scenarios. - … Find more on the tool in the core GitHub repository of the EDGELESS project: https://bit.ly/41w34Ki Abstract: https://bit.ly/41ywh7e Slides: https://bit.ly/41zjpxW

?? Our latest newsletter is here! Didn’t receive it? Not subscribed yet? No worries! Catch up on the latest news, project updates, and insights in this edition: https://bit.ly/3QlPm6s ?? Stay ahead of the game! Subscribe now to receive future newsletters directly in your inbox: https://bit.ly/3LwTcqD ?? Feel free to share with your network! #Newsletter #Innovation #StayUpdated #ProjectUpdates

?? Stay Connected with EDGELESS! Want to stay up to date with the latest in Edge Computing, AI, and Robotics? Our #EDGELESS newsletter delivers exclusive updates, project highlights, and upcoming events straight to your inbox! ?? Subscribe today : https://bit.ly/3LwTcqD #EdgeComputing #AI #Robotics #IoT #Newsletter #HorizonEU

??? Smart City Use Case: AI-Powered Urban Safety! ?? How can Computer Vision & Edge Computing improve city security? ?? The EDGELESS Smart City Use Case demonstrates real-time object and event tracking using edge nodes deployed in Smart Lamp Posts. ?? The video highlights how EDGELESS detects vehicle accidents and immediately alerts emergency responders, ensuring faster intervention by medical and police teams. ???? ?? Watch the demo and see EDGELESS in action! https://bit.ly/3BZWkdr #SmartCity #EdgeComputing #ComputerVision #AI #UrbanSafety #HorizonEU

Real-Time AI-Powered Robotics with #EDGELESS! How does the IoRT application enable seamless robot-human collaboration? By leveraging EDGELESS capabilities, robots can: ?? Understand human voice commands via Large Language Models (LLM) ?? Identify & categorize objects in real time ?? Adapt dynamically to changes in their environment With low-latency edge computing, real-time decision-making & persistent data functions, robots become smarter, safer & more efficient! Explore the details in our latest blog: https://bit.ly/42J3vlo #AI #IoRT #EdgeComputing #Robotics #Automation #SmartManufacturing #HorizonEU