Performance Improvement IoT Applications Through Multimedia Analytics Using Big Data Stream Computing Platforms

Performance Improvement IoT Applications Through Multimedia Analytics Using Big Data Stream Computing Platforms

Rizwan Patan (VIT University, India), Rajasekhara Babu M (VIT University, India) and Suresh Kallam (VIT University, India)
Copyright: © 2018 |Pages: 22
DOI: 10.4018/978-1-5225-2947-7.ch015
OnDemand PDF Download:
No Current Special Offers


A Big Data Stream Computing (BDSC) Platform handles real-time data from various applications such as risk management, marketing management and business intelligence. Now a days Internet of Things (IoT) deployment is increasing massively in all the areas. These IoTs engender real-time data for analysis. Existing BDSC is inefficient to handle Real-data stream from IoTs because the data stream from IoTs is unstructured and has inconstant velocity. So, it is challenging to handle such real-time data stream. This work proposes a framework that handles real-time data stream through device control techniques to improve the performance. The frame work includes three layers. First layer deals with Big Data platforms that handles real data streams based on area of importance. Second layer is performance layer which deals with performance issues such as low response time, and energy efficiency. The third layer is meant for Applying developed method on existing BDSC platform. The experimental results have been shown a performance improvement 20%-30% for real time data stream from IoT application.
Chapter Preview


Internet of Things

Internet of Things (IoT) is a portion of forthcoming Internet and includes many billions of ‘things’ or Internet Connected Devices (ICDs) where things can communicate, sense, potentially actuate and compute as well as have multi-modal interfaces, intelligence, virtual/physical attributes and identities. ICDs can include RFIDs, sensors, social media, smart consumer appliances, business transactions, lab instruments, etc. The idea of IoT is to allow ‘things’ to be connected anyplace, anytime, with anyone and anything, ideally using any network, any service, and any path.

IoT with chips/ sensors getting less expensive, and devices getting smarter, the world is moving toward being always on. This implies increasingly information is always being sent and got between devices. Customarily it was just cellular phones, PCs, and servers conversing with each other, Now dealing multiple features the IoT’s is shown in figure 1. the IoT’s is associated devices discussing bidirectional with each other progressively. IoT’s refers to a technology paradigm wherein ubiquitous sensors numbering in the billions will be able to monitor physical infrastructure and environment, human beings and virtual entities in real-time, process both real-time and historical observations, and take actions that improve the efficiency and reliability of systems, or the comfort and lifestyle of society. The technology building blocks for IoT have been ramping up over a decade, with research into pervasive/ubiquitous computing (Zaslavsky, 2013), and sensor networks (Chandrasekaran, Cooper, Deshpande et al., 2003) forming precursors. Recent growth in the capabilities of high-speed mobile (e.g., 2G/3G/4G) and ad hoc (e.g., Bluetooth) networks (Cornwall, 2016), smart phones, affordable sensing and crowd-sourced data collection (Data Canvas Dataset, 2016), Cloud data-centers and Big Data analytics platforms have all contributed to the current inflection point for IoT.

Figure 1.

Fusion of IoT’s


This IoT idea has recently give growing to the view of IoT Big Data applications that possibly will produce billions of data streams and Zeta byte of data to deliver the information required to support timely decision making. Some are developing IoT Big Data applications include smart manufacturing, customer sentiment analysis, emergency situations awareness, remote sensing, image processing, credit card fraud detection, and so on. IoT Big Data applications need to manage and process streaming and geographically distributed data sources for multidimensional data. All these data sources are available in the present in different locations, different formats and consistent at various self-assurance levels.

Complete Chapter List

Search this Book: