A new scheduling technique to improve data management in cloud computing

• Main Author: Khan, Nawsher
• Format: Thesis
• Language: English
• Published: 2013
• Subjects: QA76 Computer software
• Online Access: http://umpir.ump.edu.my/id/eprint/37235/1/A%20new%20scheduling%20technique%20to%20improve%20data%20management%20in%20cloud%20computing.pdf; http://umpir.ump.edu.my/id/eprint/37235/
• Institution: Universiti Malaysia Pahang

Cloud Computing is an extremely successful service oriented computing paradigm and has revolutionized, modernized, and well-developed infrastructure of computing. It is being signaled as the next-generation shift which combines the Internet and computing, as a result, users will be able to access and store software, content, and data in remote servers run by other companies or by a client. Data Management is the key factor of Cloud Computing, which is „the right data in the right place at the right time’. It is also the development and execution of architectures, policies, practices, and procedures in order to manage the information lifecycle needs of an organization in an effective manner. Scheduling techniques, which are the important part of data management, are disciplines and procedures used for distributing resources between two different parties, that is, Cloud Computing provider and Cloud Computing service user. The main purposes of scheduling algorithms, architectures, and techniques are to minimize the starvation of resources and service during the right time for using. Existing models presents the whole scheduling architecture for data transferring process, by taking in two slots. External Scheduler (ES) in one, and Local Scheduler (LS) with Data Scheduler (DS) in another slot. But new proposed scheduling architecture takes all three scheduler separately. On the base of these three separate schedulers Queue Time (QT), Execution Time (ET), and Data Transfer time (DT), also have been taken separately in data transfer time calculation. Dealing with increasing huge amount of data makes the requirement more critical for efficient accessing of data. Scheduling techniques have their major involvement in managing day-by-day increased large data in cloud environment. This research proposes a new scheduling technique to calculate the Total Completion Time (TCT) for the transfer of specific amount of data. The formula for transfer time calculation has three parameters, namely the Queue Time (QT), Execution Time (ET), and Data Transfer time (DT). All these times (intervals) are different from each other and have their own importance during calculation. In previous exist models, one of these values, either QT or DT, has been ignored by taking maximum of them. Ignoring one value means decreasing the actual consuming time. The proposed model considers each parameter separately, means giving importance to each parameter. As an outcome, the Total Transferring Time for data can be the sum of QT, ET and DT in TCT. The proposed model Total Completion Time (TCT) has been evaluated by using a single server and finite population M/M/C/*/P queuing model. There is a great impact on accuracy by taking each parameter separately in the formula. Accuracy is 85% by using 56Kbps bandwidth (BW) and number of jobs (M) taken 2, it is increased up to 92.4639% for 50 jobs. The accuracy is 98.5000%; for 2 jobs, increases up to 99.1753% for jobs 50 by using BW 512kbps. Result shows that by using M > 500, stability point (where accuracy is 100%) can be achieved. Hence new technique is more efficient when we need to transfer large amount of data. Experiments showed that the proposed model is more reliable, in terms of accuracy. The proposed model has an accurate transfer time calculation, thus Cloud Computing can present its services in a more efficient manner