SPEAR Lab is committed to open science and reproducible research. We release our research prototypes, measurement tools, and datasets to enable the community to build upon our work and validate our findings.
We welcome contributions, bug reports, and feedback from the community. Join us in building the future of edge computing and Internet infrastructure.
A novel container image distribution system that enables on-demand, file-level container partitioning for distributed machine learning workloads. 2DFS dramatically reduces container startup times and bandwidth consumption by intelligently fetching only the required files when needed.
A lightweight, hierarchical orchestration framework that enables efficient management of applications across distributed, heterogeneous edge infrastructures. Oakestra provides geographic-aware service placement, hybrid virtualization support, and seamless federation across multiple administrative domains.
A lightweight browser extension for measuring network performance metrics including bandwidth, latency, and web browsing quality. NetMet was instrumental in our large-scale Starlink performance studies, enabling measurements from real users across diverse geographic locations.
A public scanning service that continuously measures and tracks Multipath TCP (MPTCP) deployment across the Internet. MPTCP.io provides real-time statistics on MPTCP support among web servers, CDNs, and cloud services, helping researchers and practitioners understand the current state of MPTCP adoption worldwide.
Tools for characterizing distributed mobile augmented reality applications at the edge