Paper:
"Performance Modelling and Verification of Cloud-based Auto-Scaling Policies"
Alexandros Evangelidis, David Parker, Rami Bahsoon
The file data.zip contains the CPU utilisation and response time traces, gathered during our experiments on Amazon EC2. The CPU utilisation traces come in a both textual and visual form, meaning for each of the text files there is an associated screenshot captured from Amazon EC2. The naming of the files is in the following format: 1911-2000-2020. 1911 is the day and month (19th of November), 2000 is the starting time of the experiment (20:00), and 2020 is the ending time of the experiment (20:20).
The additional file AzureData.zip contains the CPU utilisation traces gathered during our experiments on Microsoft Azure. The CPU utilisation traces come in a both textual and visual form, meaning for each of the text files there is an associated screenshot captured from Microsoft Azure for the respective time period. The naming of the files follows a similar convention to the Amazon EC2 data. The prefix "cd5" before the data refers to the auto-scale actions with cool-down periods of 5 minutes.
The Apache JMeter test scripts used in the paper are here: Apache-JMeter-test-scripts.zip.
PRISM models can be found at https://github.com/alexEvangelidis/prism-cloud-models.