Governor: Operator Placement for a Unified Fog-Cloud Environment
Ankit Chaudhary, Steffen Zeuch, Volker Markl
Proceedings of the 23rd International Conference on Extending Database Technology (EDBT 2020) | March 2020

Abstract:

The processing of geo-distributed data streams is a key challenge for many Internet of Things (IoT) applications. Cloud-based SPEs process data centrally and thus require all data to be present in the cloud before processing. However, this centralized approach becomes a bottleneck for processing data from millions of geo-distributed sensors on a large scale IoT infrastructure. A new line of research extends the centralized cloud with decentralized fog devices to mitigate this bottleneck. One major challenge for an SPE in this unified fog-cloud environment is to fulfill user requirements by placing operators on fog or cloud nodes. In this demonstration, we introduce Governor, an operator placement approach for a unified fog-cloud environment. Our approach consists of the Governor placement process and Governor policies (GPs). The Governor placement process utilizes heuristic-based GPs to optimize operator placement for a user query. Using GPs, administrators can control the operator placement process to fulfill specific Service-Level-Agreement (SLA). We implement Governor in the NebulaStream Platform (NES), a data and application management system for the IoT. We showcase the im- pact of GPs on operator placement for different example queries. Our demonstration invites participants to simulate the operator placement of queries and discover their characteristics. This demonstration represents a first step towards an efficient operator placement approach for upcoming IoT infrastructures with millions of sensors and thousands of queries.

Talk:

Bibtex: