الفهرس 
CHAPTER 1 IntroductionOnly 14 pages are availabe for public view
 |  | from | 32 |  |  |
| | | from | 32 | | |
Abstract
Cloud Computing has become one of the most popular technologies in recent years. In contrast to traditional solutions, where the IT services are under proper physical, logical and personnel controls, Cloud Computing moves the application software and databases to the large data centers, where the management of the data and services may not be fully trustworthy. Current speed in the development of complex applications and growth in digital data in our lives that take part in the recycling of our work and the way we live require rapid turn for the development of technology for the communications and computing. This rapid development will facilitate the workload processing and the services provided to the consumer. Today, the cost of overseeing stockpiling hardware ranges from two to ten times the securing expense of the stockpiling hardware. We see an expanding request on advances for exchanging administration load from people to programs. Data migration and application migration are one of technologies advancements that empower registering and information stockpiling administration to be autonomic and self-overseeing. Endeavors have been trying to diminish registering expenses and thus the vast majority of them begin solidifying their IT operations and later utilizing virtualization technologies. Cloud Computing is facing numerous difficulties like virtualization, virtual machine (VM) migration, energy consumption, data security, server consolidation, etc.
Current methodologies of burden administration have researched the combination of VMs utilizing the comparable kind of workload. This research work concentrates on the investigation of administration loads in cloud environments. Administration loads help in decreasing the time required to execute errands. In this thesis, a comparison is made between some load consolidation techniques. A model is proposed to implement a technique that monitors the utility of the cloud, detect overloads, and evaluate the effects of migration of VMs from overloaded to underloaded hosts on the performance of the running applications. The proposed technique detects overload hosts in cloud computing. The overload is determined based on CPU utility in each host. Specifically, CPU utilization thresholds are set so as to keep the total CPU utilization of all VMs in each host under certain threshold. The actual host utility and the utilization threshold are then compared. If the host utility is larger than the utilization threshold then the host is considered to be overloaded. Overload detection uses two ways for computing a utilization threshold; static threshold and robust statistics. An adaptive data migration model is proposed to use when an overloaded host is detected. The model selects the best VM to migrate from the overloaded host to underloaded one to achieve load minimization across the host list. The second main contribution in this thesis addresses the challenge of ensuring reliability and balancing loads across computing nodes, managed in a decentralized service computing system.