Section: New Results

Data Stream Processing on Edge Computing

Participants : Eddy Caron, Felipe Rodrigo de Souza, Marcos Dias de Assunção, Laurent Lefèvre, Alexandre Da Silva Veith.

Latency-Aware Placement of Data Stream Analytics on Edge Computing

The interest in processing data events under stringent time constraints as they arrive has led to the emergence of architecture and engines for data stream processing. Edge computing, initially designed to minimize the latency of content delivered to mobile devices, can be used for executing certain stream processing operations. Moving opera- tors from cloud to edge, however, is challenging as operator-placement decisions must consider the application requirements and the network capabilities. We introduce strategies to create placement configurations for data stream processing applications whose operator topologies follow series parallel graphs[35]. We consider the operator characteristics and requirements to improve the response time of such appli- cations. Results show that our strategies can improve the response time in up to 50% for application graphs comprising multiple forks and joins while transferring less data and better using the resources.

Estimating Throughput of Stream Processing Applications in FoG Computing

Recent trends exploit decentralized infrastructures (e.g.. Fog computing) to deploy DSP (Data Stream Processing) applications and leverage the computational power. Fog computing overlaps some features of Cloud computing and includes others, for instance, location awareness. The operator placement problem consists of determining, within a set of distributed computing resources, the computing resources that should host and execute each operator of the DSP application, with the goal of optimizing QoS requirements of the application. The QoS requirements of the application refer to processing time, costs, throughput, etc. We propose a model to estimate the application throughput at each layer of Fog computing (Devices, Edge and Cloud) by considering a given placement solution. The estimated throughput provides a useful insight to determine the amount of physical resources to meet the QoS requirements. The model allows to identify the application bottleneck, when facing data rate variations, and provides information to self-scale in or out the DSP application.