The documents distributed by this server have been provided by the contributing authors as a means to ensure timely dissemination of scholarly and technical work on a non-commercial basis. Copyright and all rights therein are maintained by the authors or by other copyright holders, not withstanding that they have offered their works here electronically. It is understood that all persons copying this information will adhere to the terms and constraints invoked by each author's copyright. These works may not be reposted without the explicit permission of the copyright holder.
|Author:||Manisha Luthra, Boris Koldehofe|
|Kind:||In proceedings - use for conference & workshop papers|
|Book title:||Proceedings of the 20th ACM/IFIP International Middleware Conference Workshop|
|Research Area(s):||Network Mechanisms & QoS|
|Abstract:||Complex Event Processing (CEP) is a powerful paradigm that can derive meaningful insights by correlating multiple data sources, e.g., in the Internet-of-Things applications. However, these applications often require deployment across a wide variety of devices ranging from mobile devices to edge and cloud or simply put: fog infrastructure. This is not easily possible using existing programming models because of missing (i) support for deployment on heterogeneous devices and (ii) important interfaces for the deployment of CEP, e.g., for developing operator placement algorithms. In this paper, we present ProgCEP: a programming model that facilitates the development of the operator placement algorithm and its deployment in a fog computing setting. In addition, it is portable and deployable on any kind of fog infrastructure and provides dynamic scaling of resources and deployment using the operator placement algorithm. We evaluate ProgCEP on its applicability and realizability on a publicly available fog testbed involving on-site, GENI and CloudLab resources using Docker tools. To this end, we enable (i) deployment of CEP using our dockerized implementation on the aforementioned fog infrastructure in less than 50 secs (on 25 distributed resources) and (ii) easy development of operator placement algorithms in terms of minimum lines of code.|
|Full paper (pdf)|