Infrastructure resources in distributed cloud data centers (CDCs) are shared by heterogeneous applications in a high-performance and cost-effective way. Edge computing has emerged as a new paradigm to provide access to computing capacities in end devices. Yet it suffers from such problems as load imbalance, long scheduling time, and limited power of its edge nodes. Therefore, intelligent task scheduling in CDCs and edge nodes is critically important to construct energy-efficient cloud and edge computing systems. Current approaches cannot smartly minimize the total cost of CDCs, maximize their profit and improve quality of service (QoS) of tasks because of aperiodic arrival and heterogeneity of tasks. This dissertation proposes a class of energy and performance-optimized scheduling algorithms built on top of several intelligent optimization algorithms. This dissertation includes two parts, including background work, i.e., Chapters 3–6, and new contributions, i.e., Chapters 7–11.
1) Background work of this dissertation.
Chapter 3 proposes a spatial task scheduling and resource optimization method to minimize the total cost of CDCs where bandwidth prices of Internet service providers, power grid prices, and renewable energy all vary with locations. Chapter 4 presents a geography-aware task scheduling approach by considering spatial variations in CDCs to maximize the profit of their providers by intelligently scheduling tasks. Chapter 5 presents a spatio-temporal task scheduling algorithm to minimize energy cost by scheduling heterogeneous tasks among CDCs while meeting their delay constraints. Chapter 6 gives a temporal scheduling algorithm considering temporal variations of revenue, electricity prices, green energy and prices of public clouds.
2) Contributions of this dissertation.
Chapter 7 proposes a multi-objective optimization method for CDCs to maximize their profit, and minimize the average loss possibility of tasks by determining task allocation among Internet service providers, and task service rates of each CDC. A simulated annealing-based bi-objective differential evolution algorithm is proposed to obtain an approximate Pareto optimal set. A knee solution is selected to schedule tasks in a high-profit and high-quality-of-service way. Chapter 8 formulates a bi-objective constrained optimization problem, and designs a novel optimization method to cope with energy cost reduction and QoS improvement. It jointly minimizes both energy cost of CDCs, and average response time of all tasks by intelligently allocating tasks among CDCs and changing task service rate of each CDC. Chapter 9 formulates a constrained bi-objective optimization problem for joint optimization of revenue and energy cost of CDCs. It is solved with an improved multi-objective evolutionary algorithm based on decomposition. It determines a high-quality trade-off between revenue maximization and energy cost minimization by considering CDCs' spatial differences in energy cost while meeting tasks' delay constraints. Chapter 10 proposes a simulated annealing-based bees algorithm to find a close-to-optimal solution. Then, a fine-grained spatial task scheduling algorithm is designed to minimize energy cost of CDCs by allocating tasks among multiple green clouds, and specifies running speeds of their servers. Chapter 11 proposes a profit-maximized collaborative computation offloading and resource allocation algorithm to maximize the profit of systems and guarantee that response time limits of tasks are met in cloud-edge computing systems. A single-objective constrained optimization problem is solved by a proposed simulated annealing-based migrating birds optimization. This dissertation evaluates these algorithms, models and software with real-life data and proves that they improve scheduling precision and cost-effectiveness of distributed cloud and edge computing systems.
|
![[NJIT logo]](http://archives.njit.edu/vhlib/images/njit_logo.gif){kind=logo}
![[Van Houten Library logo]](images/vhl-top-image.jpg)
