SOFTWARE ARCHITECTURE DESIGNING CHALLENGES MODEL FOR INTERNET OF THINGS (IOT) SOFTWARE SYSTEM
Mr. Noor Rehman1, Dr. Abdul Wahid Khan2
Department of Computer Science
University of Science & Technology, Bannu KPK, Pakistan.
Abstract: Software Architecture is the vital part of every software to provide such mechanism which can enhance
the rule of credibility in between client and vendor side organization. The IoT software will face with
failure, if the architecture of it is not designed as desired. Elaborating the challenges faced by vendor side
organization in the begging of the software development is an integral part of this study. The goal of
Software Architecture designing Model (SADCM) for IoT software development will be used to reduce
the failure chances of any software system in future use and to make a user friendly environment system.
This will assist the vendor organization to overcome on all those challenges that create hurdles in
designing of architecture for IoT software. Systematic review of the existing literature about this study
will be analyzed to find the challenges. These challenges shall be validated by empirical approach and
finally the proposed model shall be assessed by study in software Development organization.
IoT Based Real Time Warehouse Monitoring using Sparkfun ESP8266 Thing Dev and Cayenne MyDevices
Muhammad Afzal1, Hafiz Ali Hamza Gondal2, Muhammad Bilal Arshad2, Moeed Shahid2
Abstract: As Internet of Things (IoT) field is growing very fast in this modern era, and new technologies are
coming day by day which are creating huge impact on our life as well is in industrial area. Industrial
Internet of Things (IIoT) is a subcategory of IoT. With the advancement of industry trends
warehouses are becoming essential part of industry to maintain their stocks. Every product requires
specific Environment i.e. temperature & humidity to maintain its life span. Currently industries are
using manual methods to take readings and maintain their warehouse which are not effective. Human
errors and mistakes are always possible in an environment of manual methods. These methods are not
suitable for warehouses and lead to huge loss in terms of stock which creates high impact on region’s
economy. This paper presents the real time implementation of indoor and outdoor temperature &
humidity monitoring. Our proposed system has also capability of indoor smoke monitoring of
warehouse. The study focuses on the IIoT project to monitor the warehouse indoor and outdoor
conditions in real time plus instant notification alert via SMS and e-mail according to thresholds. The
method of analysis is performed on Sparkfun ESP8266 Thing Dev Board, NodeMCU implementation
with Cayenne MyDevices Cloud. Cayenne MyDevices is a real time Cloud for IoT devices which
uses Message Queuing Telemetry Transport (MQTT) Protocol to receive and send data to and from
nodes. At hardware level nodes are using DHT22 (Temperature & Humidity) and MQ-2 (Smoke &
Combustible gases) Sensors. Nodes will send the sensor readings to cloud via Wi-Fi network. Nodes
are easily movable within the area of warehouse due to Wi-Fi mode of communication. This system
will work effectively to maintain warehouse environment easily and in a systematic way.
Experimental results proved that proposed system works efficiently in warehouse environment to
increase the life span of products.
Selection of M-Payment Business Models using Analytic Network Process
Abid Ali
Abstract: The emerging technologies continuously change the structure of business processes. The dramatic
improvements in wireless communication and mobile technologies make m-payment, a reality of today’s
world. Different organizations are approaching towards this mode of payment for their added values. The
core concept in m-payment process is the selection and application of best business model. There are
many business models but still market is searching for any dominant model. This work aims to optimize
the selection process of business models by applying the Analytic Network Process, which is one of the
MCDM approaches. According to results, the most important selection criteria are user centric
architecture and response to market trends. And on the basis of the relative importance of the given
selection criteria, the most dominant business model is collaboration model.
Critical Analysis of Six Frequently Used Classification Algorithms
Taram Nayab Shah, Muhammad Zakir Khan, Mumtaz Ali, Bilal Khan, Hammad Muhammad
Abstract: Classification is the technique used to categorize the data into a given number of classes. The main goal of classification is to identify the category to which a new data will fall under. In other words, we can say that classification is the process of generalizing data according to different instances. This paper puts together the most frequently used classifications algorithms. The algorithms include are Logistic Regression, Linear Discriminant Analysis, Naïve Bayes, Decision Tree, K-Nearest Neighbors and Support Vector Machine. These six algorithms on classification problems are the starting point to explore the classification. We applied these six algorithms on the Ionosphere and diabetes dataset for binary classification. Out of six, the decision tree has surprisingly given better results as compared to others. That is 89.46% on the Ionosphere and 77.47% on diabetes respectively.
