Incremental Stream Query Placement and Deployment in Action

Abstract

A new class of distributed stream processing engines (DSPEs) has been proposed to efficiently process large volumes of geo-distributed IoT sensor data. These systems operate on a unified sensor-edge-cloud infrastructure to offload portions of computation to IoT sensors. This allows DSPEs to reduce processing latency and achieve cost-, resource-, and energy-efficiency. However, the underlying infrastructure is highly dynamic due to the presence of mobile devices, posing significant challenges for DSPEs in performing interruption-free query execution. In this demonstration, we present the combination of incremental stream query placement (ISQP) and deployment (ISQD). Together, ISQP and ISQD enable a DSPE to perform interruption-free execution of queries in a dynamic infrastructure. We present an emulated public mobility scenario and present users with various configurations to highlight the impact of dynamic infrastructure on DSPE performance. This demonstration represents the first step towards achieving data management on a highly dynamic infrastructure.