Cloud-Based Distributed Approach for Procedural Terrain Generation with Enhanced Performance

Procedural terrain generation has already gained wide acceptance, finding applications across various domains such as video games, films, and simulations. As the demand for content size, quantity, and enhanced fidelity increases, traditional methods prove inadequate for accomplishing this task. Moreover, the planetary scale introduces further constraints on data consistency and distinctiveness. Software systems used for near-real-time procedural terrain generation of large virtual worlds must fulfill high computing requirements. This paper proposes a novel cloud-based approach to execute software agent generation requests with minimized latency. The performance is improved through vertical parallelism using Morsels and horizontal parallelism using multithreading. Apart from spreading request processing per cloud computation instance, performance is enhanced through generation request distribution using a message broker.