http://hdl.handle.net/123456789/126 Sat, 04 Apr 2026 12:03:10 GMT 2026-04-04T12:03:10Z http://hdl.handle.net/123456789/15408 SCALABILITY ANALYSIS OF BLOCK CHAIN BASED SYSTEMS Muhammad, Shahid Enroll # 02-241192-020 In recent years, block-chain technologies have gained massive momentum application domains. Block-chain is a decentralized data management technology which is speculated to be a disruptive technology that can have a drastic impact on people s lives as the Internet did. As different block-chain platforms are emerging rapidly, a firm understanding of the offerings by the adopted platform for the underlying technology along with its performance analysis is both important and challenging. Many organizations have shown interest in adopting the block-chain technology in their core systems, but scalability becomes a main concern in existing block-chain platforms. The block-chain application is stepping from its inception to full maturity and establishing itself as a part ofthe internet offuture (such as in the Internet ofThings), thus scalability of the technical challenges while having billions of devices installed worldwide with the passage over time. In this work, a comprehensive overview is presented ofa major and a popular block-chain platform, known as the Hyperledger-Fabric. The work proposes a prototype while using Docker containers as the experimental setup for deploying Hyperledger-Fabric nodes and chain-code. The work also evaluates performance ofthe Hyperledger-Fabric based block chain technologies in-terms of system scalability while considering different and scenarios. Performance evaluation can help in identifying system bottlenecks that can be further utilized to develop better solutions or optimize existing ones. A methodology is presented for evaluating performance of the block-chain platform. While using this methodology, performance analysis is done along with presenting the obtained results. The experimental results are based on varying numbers of transactions and number of nodes which reflects a detailed study of the Hyperledger-Fabric platform that may also help Hyperledger-Fabric foundation to further improve the performance oftheir platform. The experiments mainly consist oftwo cases. In Case#l, the transactions are performed by only a single node within the block-chain group ofnodes depicting a low load overthe block-chain network. In Case# I, nodes are varied from 3 to 25 nodes and the selected node performs up to 2000 transactions. Whereas in Case#2, we consider a worst scenario in which all the nodes are performing transactions on the network. For Case#2, nodes are varied from 3 to 13 nodes and transactions up to 1000. While evaluating performance, two performance parameters, namely Consensus Time and Ledger Size, mainly assessed while executing the experiments. For Case#l, over 25 nodes and while having 2000 transactions, it is observed that Ledger Size consumes a disk space of 5.3 MB and for that it takes 80.18 Minutes to complete the execution of all transactions. For Case#2, over 13 nodes and while having 1000 transactions, the Ledger Size consumes 57001.9 MB of disk space and takes 270.06 Minutes to complete the execution of all transaction. Supervised by Dr. Osama Rehman Sat, 01 Jan 2022 00:00:00 GMT http://hdl.handle.net/123456789/15408 2022-01-01T00:00:00Z http://hdl.handle.net/123456789/15409 PERFORMANCE ANALYSIS OF DEEP LEARNING APPROACH FOR CLASSIFYING DDOS ATTACK FROM BENIGN NETWORK TRAFFIC Abbas, Hafsa Enroll # 02-241192-012 Cyber security has become a great issue in this technological world. There are several types of cyber-attacks that are present, where Distributed Denial-of-Service (DDoS) is one ofthe most common attack type in the cyber world. Researchers are doing their best to find a solution to get rid of DDoS attacks. With the advancement of technology day by day, millions of people across the world are relying on the internet. People are using internet in every field of life from the very basic home task to the academics ofusers are increasing day by day, security issues are also increasing. DDoS has grown significantly than normal. DDoS attacks frequency is doubled in every year but due to COVID-19 pandemic, as everything is shifted on internet. To identify and to take measures against DDoS attacks has become a necessary task. There is a need to make a system intelligent enough to detect the difference between the legitimate request and DDoS attack request. Blocking the traffic is not a solution. It is important to develop a technique which is intelligent enough to distinguish the normal and malicious traffic. research. As the number more There are many solutions available up till now. Researchers are using different techniques to get rid ofthis problem. In this research, three different approaches are used to check which one is better for cyber security dataset. The dataset used is CICDDoS 2019 comprises of different DDoS attack types. The first approach is Machine Learning approach in which Random Forest algorithm are used. Second approach consists of ANN (Artificial Neural Network) and CNN (Convolutional Neural Network). The performance of CNN and RF is almost same. Accuracy obtained by using of all the three approaches are better. In some of the attack classification, the accuracy is increased up to 99.9%. Whereas ANN algorithm has an average performance for cybersecurity dataset. There are many anomalies occurred in the performance ofANN. The performance parameters include Accuracy, Training Time, Testing Time and Confusion Matrix. CNN takes more time in training than RF but there is a very less chance of any Supervised by Dr. Osama Rehman Sat, 01 Jan 2022 00:00:00 GMT http://hdl.handle.net/123456789/15409 2022-01-01T00:00:00Z http://hdl.handle.net/123456789/15407 DATA ANOMALY DETECTION IN MARITIME TRAFFIC MANAGEMENT USING DEEP LEARNING Mashood, Ramsha Enroll # 02-241191-015 Pandemics and the current world situation affected overall world business. The maritime industry currently depends on different vessel management systems that use Automatic Identification System (AIS) messages to manage Ship activities. AIS messages include both Static and Dynamic data, which contains information related to the position, Heading, and Ship features as AIS data is in massive volume, so the possibility of missing data and anomalous information is present. Different Machine Learning and deep learning algorithms are used to train and detect anomalies to identify those anomalies related to ship classification and ship messages. CCN, SVM, Random Forest and other deep learning algorithm are used to identify anomalies regarding the position and ship flows. While SQL, PostGre are used for storage purposes. The main goal is to identify which Database is best to store AIS data and which deep learning algorithm performs betterto identify anomalies in Ship classification and transceiver class data. This thesis proposes a solution regarding the storage of AIS data to MongoDB database and identifying anomalies in ship classification and Transceiver class using Deep Learning Algorithm. This approach includes Long Short-Term Memory (LSTM), Convolution Neural Network (CNN), and Artificial Neural Network (ANN) deep Learning algorithm for identifying anomalies in Maritime Traffic Management System. So, we can conclude that the CNN model provides the highest accuracy, approximately 81%, while other deep learning algorithms like LSTM have 72% and ANN have 79.95% accuracy. The total time consumed to get 1658575 rows takes approximately 0.04sec using MongoDB cloud Database Supervised by Dr. Sohaib Ahmed Fri, 01 Jan 2021 00:00:00 GMT http://hdl.handle.net/123456789/15407 2021-01-01T00:00:00Z http://hdl.handle.net/123456789/15399 A NEURAL NETWORK APPROACH FOR INTELLIGENT COMPETENT BASED LEARNING PATH PREDICTION IN HUMAN RESOURCE MANAGEMENT SYSTEM Qureshi, Sameed Ud DIn With the advancement of technology, data has been exponentially increasing. For this purpose, there is a need to develop such expert systems which may have the capability to deal with the variety ofthe complex problems. This research proposes an expert system that assists employees in order to recognize the pattern of employee performance throughout his/her tenure. It further helps predicting learning path for such employees. This may also help in automation ofthe complete HR process and reduction ofworkload ofHR department within the organization. In the literature related to neural networks, error correction learning algorithm is one of the algorithms that may automate human resource management system for helping organizations in order to predict employees’ performances. This predication can reduce time, provide accurate information, improvement in planning and program developments, remove language biasness and improve employees’ retentions. These advantages can directly improve an organizational culture. This culture may provide a transparency to employees for their performance evaluations and also reduce communication gap between management and employees. This culture may affect an overall progress of an organizational success. Hence, this research will evaluates the performance of an error correction learning algorithm in a human resource system. For the said purpose data set of 1470 employees have been taken from Kaggle. For this purpose, sigmoid function is used to select 123 employees for the particular criteria. This research concludes that 90% accuracy has been achieved through the use of error learning algorithm in a human resource management system. This research may facilitate management in order to identify top performers of any organization in more transparent way Supervised by Dr. Sohaib Ahmed Wed, 01 Jan 2020 00:00:00 GMT http://hdl.handle.net/123456789/15399 2020-01-01T00:00:00Z