Thumbnail Image

Thesis

Browse

20195
Reset filters
Subject: Information Technology × Start date: 2017 × Type: Thesis ×
Reset filters

Search Results

Now showing 1 - 5 of 5
  • Thumbnail Image
    ThesisItemOpen Access
    Analysis of Techniques to Retrieve Big Database
    (Punjab Agricultural University, Ludhiana, 2019) Puri, Shivani; Jain, Lokesh
    In today’s world there are large amount of data which need to be processed with big databases. From the past years, there is plethora of organisations that has adopted many types of non-relational database. The goal of this research is to implement techniques to retrieve big database for the big datasets and investigate the performance of the big database techniques on CPU utilization and high-performance computing software enabling switching the SQL database with NoSQL database. In this research mainly concerns on the new technology of NoSQL database i.e. MongoDB, HadoopDB. Performance comparison of two big data techniques is carried out. The result found that Aggregation technique consumes less execution time than MapReduce technique and more efficient with MongoDB database where as MapReduce technique has less efficient with HadoopDB. Aggregation technique also produces fine relevant information results with less CPU utilization. The result also shows that MongoDB has capability to switch SQL databases as compare to HadoopDB.
  • Thumbnail Image
    ThesisItemOpen Access
    Analysis of image forgery detection algorithms
    (Punjab Agricultural University, Ludhiana, 2019) Priyanka; Derminder Singh
    Digital images are a momentous part of today’s digital communication. It is very easy to manipulate digital images for hiding some useful information by image rendering tools such as Adobe Photoshop, Microsoft Paint etc. The common image forgery which is easy to carry out is copy-move in which some part of an image is copied and pasted on another part of the same image to hide the important information. This study proposes an algorithm to spot the copy-move forgery based on exact match block based technique. The design of interface and development of algorithm have been implemented using GNU Octave 4.2.2 as an IDE and Octave as a programming language. The algorithm works by matching the regions in image that are equivalent by matching the small blocks of size b×b. The program is tested for 45 images of mixed image file formats by considering block sizes 2, 4, 6, 8, 10, 12, 14, and 16. It is observed from the experimental results that the proposed algorithm can detect copy-move image forgery in TIF, BMP and PNG image formats only. Results reveal that as the block size increases, execution time (time taken by CPU to display output) also increases but the number of detected forged images increases till block size 10 and attains saturation thereafter. Consequently block size should be set to 10 for getting good results in terms of less execution time.
  • Thumbnail Image
    ThesisItemOpen Access
    Deployment of private cloud in campus wide network
    (Punjab Agricultural University, Ludhiana, 2019) Marken, Kunj; Jain, Lokesh
    Cloud computing in the last few years has emerged as a modern computing paradigm with its establishment. This is becoming quite popular due to the reduction in cost, scalability and improvements in the IT industry. Virtualization technology has great contribution in field of computing platform. Cloud computing allows you to use services and resources online. It is used to build and maintain the necessary infrastructure for computing resources. For deploying the cloud in an organization as a private, there are various open source platforms and OpenStack retained by the comparative study. The paper describes development of private cloud on OpenStack software by installing CentOS7 operating system for cloud deployment. Different services are installed such as web server, webmail server, FTP and SSH for communication between users within the private cloud. The results depict that hardware utilization of physical machine increases after deploying virtualization. In cloud without virtualization, hardware resources are not used efficiently because of number of users is limited but in cloud with virtualization hardware resources are utilized at maximum and used by more concurrent users. The aim of paper is to present the importance of virtualization and give the best solution for providing services in private organization.
  • Thumbnail Image
    ThesisItemRestricted
    Development of Policy Script for Intrusion Detection System
    (Punjab Agricultural University, Ludhiana, 2019) Ishwar Dayal Singh; Gupta, O.P.
    The security and management of network has become a major issue in the arena of Internet. The attacker can access different types of data i.e. personal data, bank account data, and unauthorized use of system resources in campus network. Various policies and procedures have been developed to secure the network communication over the internet by employing firewalls, encryption, and virtual private networks. On the bases of security requirements, the firewall rules are created to monitor the incoming traffic. Packet filtering technique has become a regular and inexpensive approach to secure the transfer of data over the internet and is used as a first line of defense against attacks. Strong filtering techniques in IPTABLES can be used to make a network robust in nature for securing data transfer or prevent it from attacks. Research is done, not only to safe guard the network Distributed Denial of Service (DDOS) attacks but also for the management of the network bandwidth. The proposed policy script based on the size and count of packet, blocks the attacker for a period of time. With the use of this policy, it observed that 33.8% bandwidth is always available to genuine users of the IT services.
  • Thumbnail Image
    ThesisItemRestricted
    Performance Evaluation of High Performance Computing Resources and Job Management
    (Punjab Agricultural University, Ludhiana, 2019) Karwal, Anika; Gupta, O.P.
    High Performance Computing (HPC) is a highly emerging concept in the field of computer science and technology. HPC makes the use of parallel computing to solve complex computational problems at a very high speed. Data and compute intensive applications require distinct and different resources, so it becomes utmost important to manage resources and schedule jobs accordingly. To improve operational functionality and enhance utilization of HPC many hardware and software solutions have been developed. Portable Batch System (PBS) is used to manage the resources and schedule the jobs for better utilization. Resource management and job scheduling is a major research area in high performance computing. It is used for job accounting and extensible batch job queueing. Three primary goals of PBS are queueing, scheduling and monitoring the jobs. The aim of the study is to improve the usage of resources for High Performance computation using work management and job Scheduling concepts. Optimal resource utilization and job scheduling will lead to the reduction in waiting time and high speed processing. With faster processing abilities faster results will be delivered, and thus, turnaround time will be decreased and throughput of the system increases. Hence, ultimately the performance of the HPC will be efficiently high for FCFS, PSA and RR algorithm.