Scenario: Frenquency and size of batch operation
Learn how you can control the frenquency of triggering new batch operations and the number of agents that should receive the new operation, and how you can resolve the issue of network saturation.
- Scenario
- You have 10,000 agents, and the operation is to download a 100 MB asset to each agent from EFM. The expectation is that an average agent is able to download an asset in 20-30 seconds (3-5MB/s). Assuming that EFM can handle 100MB/s network IO for this purpose, this means that roughly 20-30 parallel operations can be managed by EFM, which could easily lead to network saturation.
- Analysis
- In the following diagram, you can see that the
rollingBatchOperationsFrequencyparameter defines how often EFM triggers a new batch operation, and therollingBatchOperationsSizeparameter defines how many agents should receive the new operation.
- Solution
- A good starting point could be
rollingBatchOperationsSize=20androllingBatchOperationsFrequency=5s. In the first few iterations, the queue may become full, but it will stabilize, and the operations will proceed as expected. The expectation is that roughly 25% of the queue will be free in each iteration with a 5-second frequency.
