A Framework for Data Stream Applications in a Distributed Cloud
Matteo Nardelli · ZEUS · 2016
The ever increasing diusion of sensing and computing devices enables a new generation of data stream processing (DSP) applications that operate in a distributed Cloud environment. Despite this, most of the existing solutions, such as Apache Storm, are designed to run in a local cluster. In this paper we present our extension of Storm, which provides distributed monitoring, scheduling and management capabilities. Exploiting these new functionalities, the system can improve its perfor- mance and react to internal and external changes. Finally, we analyze open challenges of placing and adapting DSP applications.