Effective Resource Management And High Security Policies In Cloud Computing Environments

Cloud computing, the long-held dream of "computing as a utility", has opened up the new era of future computing, transformed a large part of IT industry, reshaped the purchase and use of IT software and hardware, and received considerable attention from global and lo-cal IT players, national governments, and international agencies. It is a large-scale distributed computing paradigm driven by economies of scale, in which a pool of abstracted, virtualized, dynamically-scalable, highly available, and configurable and reconfigurable computing re-sources (e.g., networks, servers, storage, applications, data, and so on) can be rapidly provi-sioned and released with minimal management effort in the state-of-the-art data centers. Ser-vices are delivered on demand to external customers over high-speed Internet with the "X as a service (XaaS)" computing architecture in a pay-as-you-go model, which is broken down into three segments:"applications","platforms", and "infrastructure". Its aims are to provide users with more flexible services in a transparent manner, and with ever cheaper and more powerful processors. User can access any service at any time, from any where, by any way in the world on demand. Similarly, IT companies with innovative ideas for new application services are no longer required to make large capital outlays in the hardware and software infrastructures. By using cloud computing platforms, they can register necessary services from the Internet and are free from the trivial task of setting up basic hardware and software infrastructures, which allows them to focus on the core aspects of their business. In computational view, cloud computing is a network of data centers and is described as a powerful, low-cost, and ener-gy-efficient approach to future computing. The data centers form what we call clouds. From a sociological standpoint on the other hand, in the cloud, applications are accessible anywhere, anytime, and storage becomes infinite for all intents and purposes. And the users can access the powerful applications, platforms, and services delivered over Internet.In current stage, the following issues in cloud computing environments need further re- search:(1) High QoS issue is one of the major obstacles for opening up the new era of the long dreamed vision of computing as a QoS, as the traditional way of resource scheduling cannot measure the cost of cloud resources accurately by reason that each of the application preference on cloud system is totally different from each other, and there may be no relation-ship between the overhead application base and the way that different task cause overhead costs of resources in cloud system. So it is need to consider high QoS issue in cloud, and hence meeting the actual high QoS demand in cloud system;(2) High availability issue is one of the major obstacles for opening up the new era of the long dreamed vision of computing as a availability, where failures are normal rather than exceptional, due to the higher data scales supporting. So it is need to consider high availability issue in cloud, and hence meeting the actual high availability demand in cloud system;(3) High energy efficiency issue is one of the major obstacles for opening up the new era of the long dreamed vision of computing as a green computing with large-scale cloud data center reach enormous number and size, because of economic, environmental, and marketing reasons. So it is need to consider high energy ef-ficiency issue in cloud with global warming and ever-rising demand for more energy, and hence meeting the actual high energy efficiency demand in cloud system;(4) High dependa-bility issue is one of the major obstacles for opening up the new era of the long dreamed vi-sion of computing as a dependability. As data center is compose of large-scale, distributed virtual nodes, it is need to improve the dependability in clouds, and make the user to select data center as a service. Dependability relate to reliability, availability, safety, integrity, and maintainability of data. So it is need to consider high dependability issue in cloud, and hence meeting the actual high dependability demand in cloud system;(5) High trust issue is one of the major obstacles for opening up the new era of the long dreamed vision of computing as a trust. As the sensitive applications and data are moved into the cloud data centers, run on vir-tual computing resources in the form of virtual machine. This unique attributes, however, poses many novel trust challenges such as accessibility vulnerabilities, virtualization vulnera-bilities, and web application vulnerabilities. These challenges relate to cloud server having physical control of data, relate to identity and credential management, relate to data verifica-tion, tempering, data loss and theft. However, the management of data and services are trust-worthy in cloud data centers. Trust as a soft social security philosophy can fight against such security threats by restricting malicious entities from participating in interactions and conse-quently offers a high secure, trustworthy, and dependable cloud computing environment.Focusing on the existing high QoS issue, high availability issue, high energy efficiency issue, high dependability issue and high trust issue in clouds, this dissertation gives a system-atic, comprehensive and in-depth research in all the five aspects. Some valuabel and innova-tional achievements have been got, and the main contributions can be summarized as follows.(1) In relate to high QoS issue, the model of cloud resource scheduling is systematically analyzed in theory, by introducing the DP mode of forecasting the scale of cloud task, and then quantifying the application preferences and the user utility of multi-dimensional QoS, the objective function of multi-dimensional QoS is given at last. Combining with rapid mul-ti-objective optimization of the immune clonal, a heuristic cloud resource scheduling algo-rithm with application preference is proposed. The non-dominated antibodies are proportion-ally immune cloned by their preference priority, which are defined by their cloud application preferences. It is beneficial to enhance the forecasting accuracy of the immune gene manipu-lation, to increase the selection pressure of the cloning operation, to reduce the search space, and to increase the speed of find the optimal solution based on the application preference. Theoretical as well as experimental results conclusively demonstrate that the algorithm can improve the performance of cloud resource scheduling in availability, load balancing devia-tion, average utility and valid time rate, and hence meeting the actual demand in cloud com-puting system.(2) In relate to high availability issue, a dynamic data replication strategy is put forward with a brief survey of replication strategy suitable for distributed computing environments. It includes:(a) analyzing and modeling the relationship between system availability and the number of replicas;(b) evaluating and identifying the popular data and triggering a replication operation when the popularity data passes a dynamic threshold;(c) calculating a suitable number of copies to meet a reasonable system byte effective rate requirement and placing rep-licas among data nodes in a balanced way;(d) designing the dynamic data replication algo-rithm in a cloud. Theoretical as well as experimental results conclusively demonstrate the ef-ficiency and effectiveness of the improve system availability brought by the proposed strategy in a clouds. (3) In relate to high energy efficiency issue, Based on the principles and semantics of green cloud, a multi-metric energy consumption model M2EC is put forward. It includes:(a) profiling the mathematical relationship between energy consumption and service level objec-tives in cloud data centers, and obtaining the conditions to meet the green service level objec-tives;(b) quantifying green cloud computing by mul-ti-metric performance, such as energy consumption, carbon emission, health effect, economy and sustainability, and ac-quiring the green cloud degree by combining all of the multi-metrics;(c) modeling green cloud by the baseline scheduling algorithm FIFO and the heuristic virtual machine scheduling algorithm M2EC, meanwhile, measuring live migration cost and managing virtual machine migration, to minimize the green cloud degree and to meet the green service level objectives; and,(d) eval-uating the green service level objectives in virtual machine scheduling by the M2EC algo-rithm within the large-scale cloud data centers, and to consider both high energy efficiency and service level objectives. Theoretical as well as experimental results conclusively demon-strate that the M2EC algorithm has high potential as it provides efficient green en-hancements and significant energy saving. It constructs green-service level objectives trade-off efficiently and effectively in clouds.(4) In relate to high dependability issue, the definition of dependability in cloud systems is given and a series of widely-used quantitative indicators are presented to evaluate the de-pendability. A systematic analysis of different redundancy strategy by theoretical formula is given. Moreover, a cloud redundancy strategy case by using system virtualization technique is established to enhance the dependability and security of heterogeneous cloud computing en-vironments. Systematic theoretical conclusively demonstrates that the higher cloud dependa-bility can be obtained by increasing the number of virtual machines and the dependability of each subsystem, the lower bound of the number of virtual machines and the dependability of each subsystem can be calculated.(5) In relate to high trust issue, Based on the properties and semantics of trust, a mul-ti-faceted and multi-phased trust model in cloud computing environments named MFMP is put forward. It includes:(a) measuring the trust based on the interactive history with a specific context of trust entities;(b) mining the malicious information by analyzing the similarity, popularity, activity, and cooperation among trust entities, and acquiring the trustworthiness recommendations;(c) modeling a multi-faceted space-decayed comprehensive evaluation method for expressing recommendation trust and a multi-phased time-decayed evaluation method for calculating direct trust; and,(d) managing trust degree change with interaction time and con-text. Theoretical as well as experimental results conclusively demonstrate that the trust model MFMP can accurately detect and prevent malicious, selfish and faulty recom-mendations, can efficiently and effectively construct and reflect the real trust relationship in clouds.