How can ML benefit Queueing? Our recent work on response time analysis in layered systems using Mixture Density Networks (MDNs) has recently appeared at IEEE MASCOTS 2021. Check out the paper here: https://bit.ly/3CVXsJJ.
Our recent work on straggler prediction and mitigation in cloud datacenters using neural networks has been accepted at the IEEE Transactions on Services Computing. Check out the paper here: https://arxiv.org/pdf/2111.10241.pdf.
We are glad to announce that our work “TauSSA, a simulator for memoryless queueing networks using tau leaping and SSA”, was presented earlier today at TOSME 2021. Visit: https://www.performance2021.deib.polimi.it/wp-content/uploads/2021/11/TOSME21_paper_9.pdf.
One of the emerging ideas at the intersection of generative neural networks and edge computing, introducing Generative Optimization Networks (GONs) for memory-efficient data generation. Our latest paper, accepted in the Workshop on ML for Systems @ NeurIPS 2021, presents GONs for anomaly detection in resource constrained edge devices with significantly lower memory footprint than traditional… Read more »
We are happy to announce that our recent research work “HUNTER: AI based Holistic Resource Management for Sustainable Cloud Computing” has been accepted for publication in Elsevier Journal of Systems and Software (https://www.sciencedirect.com/science/article/abs/pii/S0164121221002211). This is a collaborative work on sustainable cloud computing using energy efficient scheduling based on COSCO and Graph Neural Networks. Try our… Read more »
We are glad to announce that our recent work “Service Demand Distribution Estimation for Microservices Using Markovian Arrival Processes” by Runan Wang, Giuliano Casale and Antonio Filieri has been accepted in the QEST 2021 international conference. This paper proposes to estimate the service demand distribution based on measurements of traffic in-between microservices with a global… Read more »