AWS RoboMaker offers a cloud-based simulation service that enables robotics developers to run, scale, and automate simulation without managing any infrastructure. AWS RoboMaker also provides an optional integrated development environment (IDE) to develop, build, and test your robotics application all from the same cloud environment. With AWS RoboMaker, there are no upfront charges and you only pay for the AWS resources you use. You will be billed separately for each RoboMaker capability and use of AWS services provided through RoboMaker, including RoboMaker Simulation, and the RoboMaker development environment.
With AWS RoboMaker Simulation, you only pay for the time your simulation job takes to run. There are no resources to manage, no upfront costs, and you are only charged when your simulation job is in “running” status. You are charged an hourly rate based on the number of Simulation Units (SUs) required to run your simulation job. A single SU provides 1 vCPU and 2 GB of memory, and the number of SUs required to run your simulation job is determined by the maximum resource requirement between vCPU and memory. For example, if your simulation job requires 1 vCPU and 4 GB of memory to run, it will require 2 SUs. A simulation job requires a minimum of 1 SU. You are billed at an hourly SU rate in increments of 1 minute, rounded up to the nearest minute. In addition, you can optionally enable GPU access with your simulation job. You are charged an hourly rate based on GPU Unit (or GU) and each GU provides access to 1 GPU device. You are billed at an hourly GU rate in increments of one minute, rounded up to the nearest minute.
With RoboMaker Simulation WorldForge, you only pay for what you use. There are no resources to manage and no upfront costs. You can use RoboMaker WorldForge to generate a world (including the 3D assets) for use with RoboMaker Simulation, and optionally export the world to your Amazon Simple Storage Service (Amazon S3) bucket for use with other applications and services. You are charged a flat rate for each world you generate, and a flat rate for each world you export.
The number of SUs required by a simulation job scales with the complexity of your robot and simulation world. A typical ground mobility robot has one camera sensor, one Lidar sensor, and two actuators. A modestly complex simulation world, like a warehouse or a retail store, contains about 100 modeled objects. For reference, running a simulation job with a typical robot and a modestly complex simulation world as described above requires about 7 vCPUs and 14 GB of memory, or 7 SUs.
Let’s consider that your robotics application release cycle is once a month and during each release, you use RoboMaker Simulation to run 100 hours of simulations for performance testing. Consider that your robot is like a typical ground mobility robot and your simulation world is like a typical warehouse as described above. Running your simulation job will require 7 SUs.
Total SU-hours used: 100 hours/month * 7 SUs = 700 SU-hours
RoboMaker simulation price per SU per hour: $0.40
Total charges: 700 SU-hours * $0.40/SU-hour = $280
Let’s say you want to use RoboMaker Simulation WorldForge to generate 50 simulation worlds a month for regression testing of your robot’s navigation stack. You plan to export 10 percent of the generated worlds to your Amazon S3 bucket for tuning and configuration with your simulation application.
RoboMaker WorldForge generation price: $1.5 / world
Total number of worlds generated per month: 50 worlds
Total charges for world generation: 50 worlds * $1.5 / world = $75
RoboMaker WorldForge export price: $5 / world
Total number of worlds exported per month: 50 worlds * 10% = 5 worlds
Total charges for world export: 5 worlds * $5 / world = $25
Total RoboMaker Simulation charges: $75 + $25 = $100