This data set (found at https://nuigalwayie-my.sharepoint.com/personal/0102835s_universityofgalway_ie/_layouts/15/onedrive.aspx?id=%2Fpersonal%2F0102835s%5Funiversityofgalway%5Fie%2FDocuments%2FDK%20PhD%20Repo%2FServerlessBenchmarkData2020&ga=1 ) is the result of calling python based benchmarking functions on AWS Lambda and Google Cloud Functions every hour over the course of one month. These functions measure: cpu information from the /proc/cpuinfo file cpu timing statistics from the /proc/stats file request and response time of a function run time of a function (request to response) function computation start and stop function computation time (start to stop) start lag i.e. cold start (request to start) uptime of VM CPU Utilization (number of primes tested for primality) memory allocation of a function memory allocation of a VM disk I/O throughput unique VM ID unique function ID array of function IDs that executed on a particular VM
The tools used for benchmarking can be found at https://github.com/psykodan/serverless_test_functions