Scientific visualization for exascale computing is very likely to require in situ processing. Traditional simulation checkpointing and post hoc visualization will likely be unsustainable on future systems at current trends due to the growing gap between I/O bandwidth and FLOPS. As a result, the majority of simulation data may be lost if in situ visualization techniques are not deployed. In situ visualization in this paradigm will be given unprecedented access to simulation output, potentially being able to process all relevant simulation output at every simulation time step, allowing for very high temporal fidelity compared to traditional post hoc visualization. However, this access poses many challenges in terms of data management, resource management and sharing, algorithm development and design, and implementation approaches. Currently, the community primarily relies on two in situ techniques: tightly coupled (on the same resource as the simulation) and loosely coupled (not sharing resources with the simulation). Each of these approaches have positive and negative factors which affect their performance under different simulation, resource, and visualization type constraints. Meaning, that for every given visualization task, it is not generally known which method would give the best performance on every data type, architecture, and set of resource constraints. Due to the lack of research and development on this topic it is still an open research problem requiring future research.