Adaptively Compressing IoT Data on the Resource-constrained Edge
summary
Paper presents ZipMate middleware that picks compressors per-file to maximize edge→cloud throughput or compression ratio, using models and runtime resource awareness. Evaluation compresses IoT datasets (soil moisture and others) with gzip bzip2 lzma zstd on 1–4 cores and 10–80 Mbps links, shows no single compressor wins and adaptive selection cuts transfer time or increases compression by about 10–31% versus fixed choices. Code and heuristics are on GitHub; sampling-based ratio prediction is work in progress.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 0:18Outline
- 0:48IoT Ecosystem
- 1:25Challenges
- 2:24Data Compression
- 3:14Compression Challenges
- 3:52Sleepmate
- 5:01Optimal Compressor Selection
- 9:45Conclusion
related talks
39mExploring the Trend Toward the Edge
55mFog Computing and IoT Solution Architecture
41mCloud Computing: The Edge talks - long road into the Cloud
49mEnable IoT with Edge Computing and Machine Learning
20mFrom Cloud to Fog Computing and IoT
53mFull-Scale Edge Computing with Eclipse ioFog: A Case Study
01h 03mZenoh: The Edge Computing Data Fabric
29mLife on the Edge