Communication using Message Queuing Telemetry Transport (MQTT) protocol in the Internet of Things (IoT) has raised significant security challenges. A lightweight Network Intrusion detection System (NIDS) model is needed to reduce the complexity processing in power-constraint IoT modules. This paper proposed a combination of two feature selection, namely XGBoost and MaxPooling1D. This combination approach able to enhances the MQTT-enabled IoT security by effectively capture the complex pattern and extracts relevant features. The algorithm is experiment in two scenarios which is the uniflow and biflow of the MQTT dataset. The results are shown outstanding accuracy, precision, recall and F1-Score, exceeding 90%. The biflow scenario outperforms the uniflow across all performance metric, making it preferable for network data classification
This lightweight NIDS model that uses ensemble approach through hybrid feature selection shows promise in improving intrusion detection in MQTT-enabled IoT contexts. This model able to improve accuracy and security against network attacks by efficiently collecting complicated patterns and extracting important characteristics. This model is ideal for deep learning-based IDS that is customised to MQTT IoT, providing real-time monitoring and taking into account MQTT's special features.
IoT devices are considered the next technology in many areas in economy, industrial as well as daily life. The increasing popularity of these IoT devices increase the vulnerabilities and security issues concerning the IoT network. Maintaining the security has become the priority of the successful development of an IoT system. This paper introduces the integration of Software-Defiend network (SDN)-IoT platform and used Snort-based Intrusion Detection technique to detect the DDoS attack. Datasets are used as the background traffic and DDoS attack is emulated in the simulation. The adopted IDS in SDN-IoT platform is validated in hardware testbed platform. The local rule was able to be configurate according to the data of the testbed. The experimental result shows that the IDS is able to detect the abnormal traffic in the system and it is also shown using Wireshark probing.
The analysis from the testbed platform allows the local rule to be configured to enable the detect the DDoS attack. The implementation of the local rule in the SDN-IoT testbed shows that the IDS is able to detect the DDoS attack. In conclusion, the local rule for IDS configuration depends on the detection time before controller malfunction and occurrence of the attack and at the same time, security level of the IoT platform can been increased significantly
Water management is an important element in the paddy production. Water management can increase the productivity level, thereby increase the income of farmers. Traditional rice is grown under continuous submergence or intermittent or variable ponding conditions depending on the farmer’s choice and also on the available water resources. The paddy crops are sensitive to the acidity of the soil and water. To enable analysis of water quality and the soil in different paddy bunds at during stages of paddy cultivation datasets are very important. Environment, water quality and soil quality parameters need to be taken continuously and enormous data will be collected. In order to provide meaningful data to Department of Agriculture, Department of Irrigation and Drainage or even the farmers, the data need to be analyze and characterize. The project will be conducted in three phases which are outdoor dataset collection, the data pre processing and analyze, the characterization of predictive analysis. The outcome from this project will benefit the department of agriculture, department of irrigation and drainage, and the paddy farmers in terms of paddy cultivation and water quality at the paddy site. This will uphold the National Agrofood Policy that to increase the income of the agropreneurs.
Researchers working on water quality monitoring usually need to go to a location to take water samples. These outings require cost and time where sometimes the location is very remote. Water monitoring is popular for cage aquaculture in the inner sea and fresh water. There is also water monitoring for lake resource management. Many parts of the world such as Korea, China, Japan as well as Malaysia are trying to increase the production of aquaculture products that are environmentally friendly. Environmental friendly aquaculture is different to the one offered for agricultural and yet aquaculture has many enemies such as fish diseases, humidity, and pollution. Thus there are needs for water quality monitoring to improve the production processes such as breeding data, food, medication, vaccine or some logistic information. WQ1.0 is a module that is able to monitor the quality of the water and transmit the data wirelessly to the cloud. These data can be monitor from the office or any mobile devices. This module is the output of PRGS grant under MOSTI 2014-2016.
Field test on seahorse hachery area
WQ1.0 has been verified with commercial water monitoring device