AWS recently released SimSpace Weaver, a managed option to run real-time spatial simulations across multiple EC2 instances. Distributing simulation workloads, the service can handle large real-world environments, crowd simulations, and immersive interactive experiences.
SimSpace Weaver divides the virtual world spatially and distributes the pieces across a cluster of compute instances. The new service manages the network and memory configurations, synchronizing data across the instances to create a single simulation where concurrent users can interact and manipulate it in real-time.
Source: https://aws.amazon.com/blogs/aws/new-aws-simspace-weaver-build-large-scale-spatial-simulations-in-the-cloud/
Among the possible use cases, the cloud provider includes large-scale crowds, creating millions of dynamic entities with unique behavior models, or city-scale environments, simulating traffic patterns or public transportation networks. As for the SimSpace Weaver and uCrowds city demo, the service can create immersive and interactive experiences, preparing teams to respond to real-world scenarios by building training environments. Marcia Villalba, principal developer advocate at AWS, writes:
Organizations run simulations on situations that are rare, dangerous, or very expensive to test in the real world. For example, city managers can’t wait for a natural disaster to hit a city to test the response systems. Event planners don’t want to wait until a large sporting event to start to understand the impact the games will have on traffic. Scenarios like these need to be simulated in a safe environment in which planners can test different situations and tune each system.
Developers define the boundaries of a simulation world and divide it into a grid. Instead of creating simulation logic that works on the entire grid, they can focus on one grid cell operating a spatial app that includes the logic for all entities in that space. Villalba highlights how the service supports engines like Unity and Unreal Engine 5:
SimSpace Weaver provides the infrastructure to weave a simulation across multiple instances, but it is not a simulator (...) Developers building with Unreal Engine 5 or Unity can take advantage of the SimSpace Weaver out-of-the-box plugins to jump-start their projects.
AWS provides a quick start tutorial to build and run a simulation and SimSpace Weaver Local, an option to configure and test a simulation locally before uploading it to the AWS service. SimSpace Weaver has two different sets of APIs, Service APIs to control the main resources, such as simulations, clocks, and apps, and App SDK APIs to manage the data within the simulations, for example, read and write entity field values.
Yasunori Kirimoto, geospatial architect and AWS DevTools Hero, documents the steps of a simple simulation using the new service.
Developers are charged for the number of EC2 instances used for the length of their simulation. The service currently supports up to ten sim.c5.24xlarge instances, with 17 application slots per instance, starting at 6.08 USD per instance per hour.
SimSpace Weaver is available in a subset of AWS regions, including Ohio, Northern Virginia, Ireland, Frankfurt, and Stockholm.