The Benchmarking Suite is composed by five main components summarized in the following diagram:
- Core: defines the main model, the extension framework for the benchmarks and the Cloud providers and the main data representation types;
- REST: a REST server to access all the Benchmarking Suite functionalities;
- CLI: a command line tool (bash-compatible) to access all the Benchmarking Suite functionalities;
- Std Library: a set of selected benchmark tools, including their configurations and the implementation of the required wrapping scripts for the execution and the parsing of results;
- Backend Connectors: a set of connectors to store the Benchmarking Suite executions results on different storage technologies (e.g. MySQL, MongoDB).
- Scheduler: a service that periodically executes a benchmarking session;
The Core component is the only required component, the other components are optional. However the Benchmarking Suite installation will miss the functionalities of not-installed modules (e.g. if the Backend Connectors is not installed, the execution results will not be stored).
The User’s Cloud Configuration is the required configuration of the Cloud Providers that the Benchmarking Suite needs to be able to access the Target Cloud Provider. It can be specified either as configuration file or as parameter in the execution requests (through the REST or CLI components). Refer to section Service Providers for further details
The core concept in the Benchmarking Suite is the BenchmarkExecution. It represents the execution of a Benchmark test against an ExecutionEnvironment provided from a ServiceProvider and produces an ExecutionResult.
For instance, following this model we can easily model the execution of YCSB.WorkloadA (the Benchmark) on the Virtual Machine with ip=22.214.171.124 (the ExecutionEnvironment) provided by Amazon EC2 (the ServiceProvider).
Since it is frequent to execute multiple tests against the same Service Provider, the Benchmarking Suite has also the concept of BenchmarkingSession. that can gropu one or more executions of the ServiceProvider, using the same ExecutionEnvironment.
In order to address all the different use cases and the installation necessities, the Benchmarking Suite is distributed in six different software modules that can be installed separately:
||the core library (all other modules depend on it) with the definition of types and the fundamental framework for the extension of the Benchmarking Suite|
||a collection of benchmark tests configuration files and support for some Cloud Providers|
||a bash command line tool to manage tests and results|
||an HTTP server and a REST API to interact with the Benchmarking Suite|
||connectors for the supported storage backends|
||a service to automatically execute periodic benchmarks|