Agent-based modeling, or agent-based model (ABM), is a powerful tool that helps researchers understand complex and high-stakes problems, such as the impact of climate change, digital twins, or epidemiology issues. However, as the scale and/or the precision of these models increase, performance problems emerge when running large simulations on standard computers. Agent-based simulation platforms, can no longer run these large-scale models in a reasonable amount of time, or even cannot run them at all. To address these problems, ABM can be distributed using High-Performance Computing (HPC) techniques to divide the workload across multiple processors, speeding up the simulation execution. Distribution bring the need for defining data synchronization protocols between processors, and ways to access agents across processors to properly execute the distributed simulation. To tackle the issues introduced by the distribution of a model, we propose the concept of proxy: entities managing interactions between agents across distributed instances of a simulation and controlling access to agent data using data synchronization policies. The effectiveness is shown through a case study using the GAMA platform.