SOF: Zero Configuration Simulation Optimization Framework on the Cloud

Simulation models are becoming an increasingly popular tool for the analysis and optimization of complex real systems in different fields. Finding an optimal system design requires performing a large parameter sweep. Hence, the model tuning process is extremely demanding from a computational point of view, as it requires careful, time-consuming, complex orchestration of coordinated executions. In this paper, we present the design of SOF (Simulation Optimization and exploration framework on the cloud), a framework which exploits the computing power of a cloud computational environment in order to realize effective and efficient simulation optimization strategies.

SOF offers several attractive features: firstly, SOF requires “zero configuration” as it does not require any additional software installed on the remote node (only standard Apache Hadoop and a SSH access are sufficient). Secondly, SOF is transparent to the user, since the user is totally unaware that system operates on a distributed environment. Finally, SOF is highly customizable and programmable, since it enables the running of different simulation toolkits and/or the ability to exploit diverse programming languages – provided that the hosting platform support them – under two different simulation optimization scenarios, as developed by the modeler.

The framework core has been fully developed and is available under the Apache public licence. It has been tested and validated on several private platforms, such as a dedicated cluster of workstations, as well as on public platforms, including the Hortonworks Data Platform (Hortonworks).

SOF was designed in ISISLab and allows the simulation modeller to run and collect results in two kinds of scenario parameter space exploration (PSE) and simulation optimization (SO) considering the computational resources as available for a not fixed time and subjects to failure.

SOF was designed to manage three kinds of simulation engine: MASON, NetLogo and a generic simulator. SOF provides some software facilities for the first simulators like the automatic simulation input setting and automatic output generating (that does not provide for the generic simulator, for obvious reasons). The generic simulator must be an executable compliant with the cluster machine used.

SOF is a framework to exploit simulation optimization on Hadoop cluster. SOF is divided in two main functional blocks: core and client. The core component provides all functionality to write out Java based client application. The client is a command line Java application that shown the features of the core component and allows to execute PSE and SO process on a Apache Hadoop cluster.

Download Star