Companion of the SimGrid storage modeling article
The tremendous increase in scientific data production and the ever-growing need for data analysis and preservation coming from various scientific domains create a great emphasis on storage components. Understanding the performance of a storage subsystem or dimensioning it properly is an important concern independent of the scale and type of the associate computing infrastructure. Data centers, Supercomputers, Grids, and Clouds comprise storage components whose specifics may differ but all need to be well understood. The study of such systems often mandates empirical evaluation of solutions via simulation. Unlike direct experimentation, simulation enables fully repeatable and configurable experiments that can often be conducted quickly for arbitrary hypothetical scenarios. However, most simulation frameworks tailored for the study of distributed systems offer no or little abstractions of storage infrastructures.
In this paper, we present our efforts to extend the simulation capacities of the SimGrid framework with abstractions, models, and interfaces that allow users to manipulate storage components and their contents in their simulations. We detail our design choices, the proposed implementation, and experiments designed to study the validity of the proposed model.