FTR: Performance-Aware and Energy-Efficient Communication Protocol for Integrating Sensor Networks into the Internet

Integrating sensor networks into the Internet brings many advantages. For example, users can monitor or control the state of the sensors remotely without visiting the field. Some researchers have proposed methods using a REST-based web service or HTTP to establish communication between sensors and s...

Full description

Bibliographic Details
Main Authors: Sinung Suakanto, Suhono H. Supangkat, S. Suhardi, Roberd Saragih
Format: Article
Language:English
Published: ITB Journal Publisher 2014-11-01
Series:Journal of ICT Research and Applications
Online Access:http://journals.itb.ac.id/index.php/jictra/article/view/833
Description
Summary:Integrating sensor networks into the Internet brings many advantages. For example, users can monitor or control the state of the sensors remotely without visiting the field. Some researchers have proposed methods using a REST-based web service or HTTP to establish communication between sensors and server via the Internet. Unfortunately, as we know, HTTP is a best-effort service. In some cases this means that if the number of sensors increases the end-to-end Quality of Service will decrease. The end-to-end network delay increases, as well as the failure rate of data sending caused by HTTP timeouts. In this paper, we propose Finite Time Response (FTR) HTTP as a communication protocol suitable for integrating sensor networks into the Internet. We have defined a cross-layer approach that coordinates between the application layer and the physical layer to control not only performance but also energy efficiency. The HTTP request-response delay measured at the application layer is used as the decision factor at the physical layer to control the active and sleep periods. We also propose a forced-sleep period as a control mechanism to guarantee average performance for all nodes. The experimental results have shown that FTR has the ability to maintain better performance, indicated by a lower average response time and a lower average timeout experience. Optimization is still needed to gain better performance and better energy efficiency while also considering the average value of the update time.
ISSN:2337-5787
2338-5499