Edge AI workloads demand low-latency, privacy-compliant storage. CATER delivers, using policy-driven placement to optimize energy, latency, and compliance. Tested on real edge clusters, it cuts data movement by 61%. 🌟 Follow us on LinkedIn! https://www.linkedin.com/company/clever-project/?viewAsMember=true🌟 Check the updates from the website: www.cleverproject.eu 🌟 Full paper in: https://zenodo.org/records/11029807

Combining LSTM for QoS prediction and LOF for anomaly detection, our predictive retraining adapts to real-time traffic changes in B5G networks. The result? Fewer violations, smarter resource use, and 99% accuracy in detecting new patterns. 🌟 Follow us on LinkedIn! https://www.linkedin.com/company/clever-project/?viewAsMember=true🌟 Check the updates from the website: www.cleverproject.eu 🌟 Full paper in: https://zenodo.org/records/11045683

Static retraining fails in dynamic 6G networks. Our predictive approach uses LOF and LSTM to detect traffic shifts and retrain just in time, reducing SLA violations by 40% compared to threshold-based methods. 🌟 Follow us on LinkedIn! https://www.linkedin.com/company/clever-project/?viewAsMember=true🌟 Check the updates from the website: www.cleverproject.eu 🌟 Full paper in: https://zenodo.org/records/11045683

Apache Ozone excels at scalable storage—but lacks constraint-aware placement. CATER fills the gap, integrating seamlessly to reduce active nodes by 23% while respecting GDPR, hardware, and collocation rules. 🌟 Follow us on LinkedIn! https://www.linkedin.com/company/clever-project/?viewAsMember=true🌟 Check the updates from the website: www.cleverproject.eu 🌟 Full paper in: https://zenodo.org/records/11029807

RDMA’s memory registration can kill performance—unless you pre-register buffers. Our GVirtuS RDMA communicator does exactly that, slashing context switches by 98% and boosting GPGPU virtualization speed by up to 55%. 🌟 Follow us on LinkedIn! https://www.linkedin.com/company/clever-project/?viewAsMember=true🌟 Check the updates from the website: www.cleverproject.eu 🌟 Full paper in: https://zenodo.org/records/14717622