vHive Serverless Function Snapshot Performance Benchmarks

Benchmark data from the vHive open-source framework characterizes snapshot-based serverless infrastructure using Containerd and Firecracker. The analysis reveals a 95% average increase in execution time for functions started from snapshots versus memory-resident ones and demonstrates a 3.7x reduction in cold-start delays with the REAP prefetching mechanism. The work was conducted by Dmitrii Ustiugov of the University of Edinburgh.

Use Cases

• Benchmarking cold-start latency in serverless platforms based on snapshotting techniques described in the paper.
• Analyzing page fault patterns and memory working sets for serverless function invocations.
• Evaluating the performance impact of prefetching mechanisms like REAP on serverless cold starts.
• Studying the trade-offs between snapshot-based and memory-resident serverless function execution.

Strengths

• Analysis is based on a characterized state-of-the-art snapshot infrastructure using industry-leading technologies (Containerd, Firecracker).
• Quantifies a 95% average execution time increase for snapshot-started functions and a 3.7x average cold-start delay reduction with REAP.

Limitations

• Row count and dataset size are unknown, which may limit suitability assessment.
• Column-level documentation is absent; field semantics must be inferred after download.
• Last update date is unknown; freshness unverified.

Provenance

Source

University of Edinburgh (Dmitrii Ustiugov) via paperswithcode.

Collection Method

Likely contains experimental results and benchmarks generated using the vHive framework.