More than bin packing: Dynamic resource allocation strategies in cloud data centers
CSV files with aggregated statistics of our simulations and experiments as well as R scripts that we used to generate all tables found in the paper. Additionally it contains the schedules describing the allocation and deallocation of VMs.
Time series data used by our simulations and experiments to describe a varying workload scenario for each VM. Each VM arrival and departure schedule references one or multiple time series found in this package. Time series are encoded using Apache Thrift and can be accessed using the Times service.
RAW monitoring data collected in our testbed infrastructure during all experiments reported in this paper. There are multiple CSV files, each describing the time series for one host and sensor (CPU, memory, disk utilization) combination. Furthermore there are CSV files describing the log output of our controller implementations and workload driver (e.g. Rain or Cloudburst). Originally the data was collected using the Sonar monitoring system which was used to monitor our whole testbed infrastructure.
HDFS (Hadoop Distributed File System) dump of our Sonar monitoring database. This data was used by all our analysis scripts to generate the aggregated CSV files found in main.zip. The package also covers additional data that was used in multiple earlier experiments.