An engineering firm is running a scientific workflow powered by two tightly integrated applications. One is a Linux-based simulation engine that writes output data to an NFS file share every 5 minutes. The other is a Windows-based desktop visualization tool that reads from an SMB file system to display the simulation results. Currently, the team maintains two mirrored file systems to support both applications, leading to data duplication and inefficient use of infrastructure. The company wants to migrate this system to AWS but cannot make changes to either application. What is the best solution that meets these constraints?