Paper accepted at EuroSys’25
Our paper “CAPSys: Contention-aware task placement for data stream processing” was accepted at EuroSys’25!
We performed an empirical evaluation study to show that task placement not only significantly affects streaming query performance but also the convergence and accuracy of auto-scaling controllers. To address this issue, we propose CAPSys, an adaptive resource controller for dataflow stream processors, that considers auto-scaling and task placement in concert. CAPSys relies on Contention-Aware Placement Search (CAPS), a new placement strategy that ensures compute-intensive, I/O-intensive, and network-intensive tasks are balanced across available resources.
You can read the technical report here.