GHTraffic: A Dataset for Reproducible Research in Service-Oriented Computing

We present GHTraffic, a dataset of significant size comprising HTTP transactions extracted from GitHub data (i.e., from 04 August 2015 GHTorrent issues snapshot) and augmented with synthetic transaction data. This dataset facilitates reproducible research on many aspects of service-oriented computing.

The GHTraffic dataset comprises three different editions: Small (S), Medium (M) and Large (L). The S dataset includes HTTP transaction records created from google/guava repository. Guava is a popular Java library containing utilities and data structures. The M dataset includes records from the npm/npm project.  It is the popular de-facto standard package manager for JavaScript. The L dataset contains data that were created by selecting eight repositories containing large and very active projects, including twbs/bootstrap, symfony/symfony, docker/docker, Homebrew/homebrew, rust-lang/rust, kubernetes/kubernetes, rails/rails, and angular/angular.js. 

We also provide access to the scripts used to generate GHTraffic. Using these scripts, users can modify the configuration properties in the config.properties file in order to create a customised version of GHTraffic datasets for their own use. The readme.md file included in the distribution provides further information on how to build the code and run the scripts.

The GHTraffic scripts can be accessed by downloading the pre-configured VirtualBox image or by cloning the repository.