Resource-Efficient DNN Inference with Early Exiting in Serverless Edge Computing

X Guo, F Dong, D Shen, Z Huang… - IEEE Transactions on …, 2024 - ieeexplore.ieee.org
Serverless Edge Computing (SEC) has gained widespread adoption in improving resource
utilization due to its triggered event-driven model. However, deploying deep neural network …

Coach: Exploiting Temporal Patterns for All-Resource Oversubscription in Cloud Platforms

B Reidys, P Zardoshti, Í Goiri, C Irvene… - arxiv preprint arxiv …, 2025 - arxiv.org
Cloud platforms remain underutilized despite multiple proposals to improve their utilization
(eg, disaggregation, harvesting, and oversubscription). Our characterization of the resource …

PISeL: Pipelining DNN Inference for Serverless Computing

M Rahimi Jafari, J Su, Y Zhang, O Wang… - Proceedings of the 33rd …, 2024 - dl.acm.org
Serverless computing offers resource efficiency, cost efficiency, and a" pay-as-you-go"
pricing model, which makes it highly attractive to both users and cloud providers. However …

Fast and Efficient Memory Reclamation For Serverless MicroVMs

OL Nikolos, C Alverti, S Psomadakis, G Goumas… - arxiv preprint arxiv …, 2024 - arxiv.org
Resource elasticity is one of the key defining characteristics of the Function-as-a-Service
(FaaS) serverless computing paradigm. In order to provide strong multi-tenant isolation …

[PDF][PDF] CXLfork: Fast Remote Fork over CXL Fabrics

C Alverti, S Psomadakis, B Ocalan, S Jaiswal, T Xu… - 2025 - tianyin.github.io
The shared and distributed memory capabilities of the emerging Compute Express Link
(CXL) interconnect urge us to rethink the traditional interfaces of system software. In this …

Towards microarchitectural side-channel security for modern applications—a case for many-domain processors

S Wei - 2024 - repositories.lib.utexas.edu
Modern applications often host or process code and data from many mutually distrusted
entities referred to as security domains. Ensuring the data privacy and security of these …