Comparison of cold starts for different serverless providers

Interesting article which compares how long different functions-as-a-service back-ends perform when starting up a server to handle a function request after a long idle period with no calls: https://mikhail.io/2018/08/serverless-cold-start-war/

Comparisons are done not only between AWS, Google, and Azure, but also looking at different languages for the functions, different memory requirements and even different numbers of packages for JavaScript functions. Good comparison!