This repository contains code, scripts, and results for benchmarking the execution of lambda functions.
This repository depends on you having a reasonable terraform, ruby, and docker environment available on your system. The ruby environment should include an installation of the v2 aws-sdk gem.
To run the benchmarks, first you must build the lambda deployment zips.
This can be done by typing make and waiting a while.
It should complete with no errors.
After that completes, you should run terraform init, followed by ./run.sh.
All together:
$ make # should take a while to pull docker images and build
$ terraform init
$ AWS_PROFILE=my-aws-profile ./run.sh
# Wait *3 hours*
# Do not run `./run.sh` in parallel on one account; use multiple AWS accounts
# if you want to make things go quicker
# data should exist in `./output` nowThe results I got have been committed to this repo, along with some python code
to plot them. This code is in ./util/process_results/process_results.py, and
running it should be fairly trivial.