An SDN-Assisted Framework for Optimal Deployment of MapReduce Functions in WSNs

In this paper, we propose a solution to support big data processing with MapReduce inside a wireless sensor network (WSN). The proposed framework provides a simple MapReduce interface, whose implementation can be dynamically loaded into the appropriate nodes. The execution of map and reduce function...

Full description

Saved in:
Bibliographic Details
Published in:IEEE transactions on mobile computing 2016-09, Vol.15 (9), p.2165-2178
Main Authors: Anadiotis, Angelos-Christos G., Morabito, Giacomo, Palazzo, Sergio
Format: Magazinearticle
Language:English
Subjects:
Big data
Computer architecture
Context
MapReduce
Mobile computing
Network topologies
Network topology
Sensors
Software
software defined networking
Wireless sensor networks
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In this paper, we propose a solution to support big data processing with MapReduce inside a wireless sensor network (WSN). The proposed framework provides a simple MapReduce interface, whose implementation can be dynamically loaded into the appropriate nodes. The execution of map and reduce functions inside the network is achieved by leveraging software defined networking principles and in particular by enhancing the SDN-WISE protocol on both the Controller and the sensor node sides. On the Controller side, the proposed framework introduces a functionality, which evaluates the optimal deployment of the processing functions in the network taking the typical WSN requirements into account and enforces it by instructing the flow tables to forward traffic from the mappers to the appropriate reducers. Then, on the node side, SDN-WISE is integrated with Contiki, a widely popular operating system for sensor nodes, which has the capability to load and execute new functions without service disruption, i.e., on the fly. The proposed framework is evaluated considering different network topologies and approaches about the data processing location.
ISSN:1536-1233
1558-0660
DOI:10.1109/TMC.2015.2496582