Please use this identifier to cite or link to this item: http://hdl.handle.net/10603/292583
Title: A Novel Hybrid Approach for Load Balancing and Resource Aware Optimization in Cloud Environment
Researcher: Dinesh Gupta
Guide(s): Harmaninderjit Singh
Keywords: Computer Science
Computer Science Information Systems
Engineering and Technology
University: Desh Bhagat University
Completed Date: 8.2.2020
Abstract: newline The term cloud comes from the symbol used to represent internet or network. It is gaining lot of popularity as it provides low cost access to high computing resources and extremely large storage spaces with high level security. Cloud computing refers to many different types of services and applications being delivered over the internet cloud. Cloud computing provides an efficient and affordable means of providing platforms such as IaaS (Infrastructure as a service), SaaS (Software as a service), PaaS (Platform as a service) and much more to the service requesters. A large number of commercial applications are available that provide wide range of services over cloud. Increased and fast access to internet has also led to a boom in this sector. Different services are put within the virtualized resources of a cloud, enabling it to carry out abstractions of its underlying resources. Cloud computing environment provides virtualized resources to applications dynamically. These resources are costly and often impossible for users to own them privately. Users are charged on a pay-per-use basis for these resources on cloud environment. newlineScheduling of tasks in cloud environment is a key research area. The problem involves allocation of almost infinite number of tasks on cloud resources that cannot be solved in polynomial time and hence is considered to be a NP-hard problem. The solution provided to close to optimal as no algorithm exists that provides optimal solution in finite time. There are no algorithms which may produce optimal solution within polynomial time to solve these problems. The idea of task scheduling in cloud environment is to optimize the execution time so that the cost paid by the user is minimal. Task scheduling is the process of distributing workloads across multiple computing resources. Load balancing is an optimization problem and goal of any optimization is to either minimize effort or to maximize benefit. The effort or the benefit can be usually expressed as a function of certain design variables. Hence, optimization is the process of finding the conditions that give the maximum or the minimum value of a function. Load balancing is a problem where you try to minimize value of parameters like Makespan time, Response Time, etc. and increase the utilization of cloud resources. newlineIn this thesis, performance of different static and stochastic meta-heuristic algorithms is compared with the proposed algorithm. Static algorithms are easy to implement but they fail to provide even acceptable solutions. Ant based algorithms are very popular for task scheduling related problems in cloud computing.Autonomous agent based load balancing algorithm (A2LB) is a dynamic ant based resource scheduling algorithm that provides scalability and reliability by offering better resource utilization, and minimum response time, but it results in high degree of migration. Particle Swarm Optimization (PSO) is a swarm-based meta-heuristic algorithm influenced by the social behaviour of animals such as bird or fish. PSO has fewer primitive mathematical operators than other metaheuristic algorithms which results in lesser convergence time and is applied to continuous value problems. Ant Colony Optimization (ACO) is also a swarm based meta-heuristic algorithminspired by the behaviour of real ants looking for the shortest path between their coloniesand a source of food. Ant Colony Optimization algorithm is suited for solving discrete problems and can be used in solving the cloud computing resource management and job scheduling.Algorithm proposed in thesis tries to overcome the limitations of ACO and PSO. The thesis is organized as follows: newline newlineChapter 1 Gives a brief introduction about cloud computing. It gives an insight into the evolution of operating system followed by various service models and deployment models used in cloud computing. Then it discusses various challenges in cloud computing. newlineChapter 2 Explain in detail the architecture of cloud computing and how visualization is implemented to make effective use of cloud computing concept. newlineChapter 3Describes the concept of optimization and gives an overview of different concepts related to optimization. It also gives introduction to ant colony optimization and particle swarm optimization techniques. newlineChapter 4Gives an introduction about cloud scheduling problem and various constraints and metrices that are involved in while designing solution to task scheduling in cloud. newlineChapter 5Gives a detailed review of the literature of work already done in the field of task scheduling in cloud. newlineChapter 6Discusses the results. It provides in depth comparison between existing scheduling algorithms and proposed scheduling algorithm. newlineChapter 7Gives conclusion and future work that can be done in the field of task scheduling in cloud computing. newline
Pagination: 
URI: http://hdl.handle.net/10603/292583
Appears in Departments:Department of Engineering and Technology

Files in This Item:
File Description SizeFormat 
80_recommendation.pdfAttached File112.65 kBAdobe PDFView/Open
a novel hybrid approach for load balancing and resource.pdf90.54 kBAdobe PDFView/Open
certificate of declaration by candidate.pdf59.83 kBAdobe PDFView/Open
chapter 1.pdf154.47 kBAdobe PDFView/Open
chapter 2.pdf386.1 kBAdobe PDFView/Open
chapter 3.pdf732.26 kBAdobe PDFView/Open
chapter4.pdf473.33 kBAdobe PDFView/Open
chapter 5.pdf116.56 kBAdobe PDFView/Open
chapter 6.pdf600.51 kBAdobe PDFView/Open
chapter 7.pdf544.16 kBAdobe PDFView/Open
preliminary data.pdf208.52 kBAdobe PDFView/Open
Show full item record


Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).

Altmetric Badge: