IIT Jodhpur collaboratively develops a novel framework to enhance the performance of IoT systems.

Modern research in the field of the Internet of Things has been conducted by scientists from the Indian Institute of Technology (IIT) Jodhpur, the Indian Institute of Information Technology Guwahati, and IIT Kharagpur (IoT). To increase the effectiveness of data collecting and transmission connected with IoT devices and apps, the team has created designs and algorithms.

The research study was co-authored by Suchetana Chakraborty, an assistant professor in the department of computer science and engineering at IIT Jodhpur, Sandip Chakraborty, an associate professor in the department of computer science and engineering at IIT Kharagpur, and Anirban Das, a research scholar in the department of computer science and engineering at IIIT Guwahati. It was published in the journal Future Generation Computer Systems, published by Elsevier.

IIT Jodhpur assistant professor Suchetana Chakraborty explained the significance of the study’s findings by stating that the Internet of Things (IoT) is the next Industrial Revolution since it is gradually transforming our way of life. Smart homes are already a reality, and with improvements in artificial intelligence, IoT systems are enabling functional robots and self-driving automobiles, among other things. We have already started connecting everyday objects to the internet via embedded devices.

Data is transferred between systems and objects in IoT systems over the internet. Currently, such data management and transmission are packed into separate ecosystems. For instance, IoT systems that operate on devices running a single operating system are unable to communicate with devices running different operating systems. IoT service sharing is becoming more popular across ecosystems, according to Chakraborty, who also noted that “such an architecture raises a basic question: how can numerous apps utilise and govern a single IoT setup?”

Data is exchanged between an end device and a processing centre in the current Internet of Things applications, which may be located on an edge server or in the cloud (the area where a device or local network interfaces with the internet). The need to transport huge amounts of data is the current issue. Although data compression techniques are employed, they do not consider whether the data is pertinent to the particular IoT application. Additionally, there is resource waste because each ecosystem is installed on its own cloud/edge server and operates independently.

The primary researcher stated, “We intended to address the above two concerns of resource loss and data irrelevance by the creation of unique algorithms. For effective data management and forwarding on shared IoT infrastructure, the team has created CaDGen (Context-aware Data Generation), an extremely edge-based data pre-processing framework. Two modules make up CaDGen. Based on the context of the active applications using the sensing infrastructure, the adaptive sensing module filters the data. The selective forwarding module selects the data forwarding paths so that various microservices running over the edge devices can utilise the data to the best of their abilities based on their needs.

According to the institute, the researchers assessed CaDGen’s performance under various conditions and found encouraging results in terms of network resource use, scalability, energy conservation, and distribution of computation for the best service providing.

For a somewhat dynamic scenario, the context analysis method could reduce the generated data by about 35% without sacrificing the data quality by filtering out information that was not relevant to the programme that was running, it continued. The researchers wrote in their paper outlining their findings, “We believe that such an approach can suit diverse smart settings in a connected living setup that lowers the expense of data storage while delivering an effective service architecture for end-users.