Orleans - Resource-based placement with cooperative dual-mode Kalman filtering
published on 2024/01/31
Orleans ensures that when a grain call is made there is an instance of that grain available in memory on some server in the cluster to handle the request. This is called "grain placement". There are a set of out-of-the-box placement strategies that developers can make use of, such as:
For the once in the audience that are not aware, I've recently published a package which performs grain placement in an attempt to achieve approximately even load, based on cluster resources. It assigns weights to silo runtime statistics to prioritize different properties (cpu usage, memory usage, available memory) and calculates a normalized score for each silo. The silo with the lowest score is chosen for placing the activation. Normalization ensures that each property contributes proportionally to the overall score. The users can adjust the weights based on their specific requirements and priorities for load-balancing.