Research On Task Scheduling In The Cloud Computing

Cloud computing as the primary means of solving big data and distributed computing provides cluster resources to the users as like services. These resources are used as needed and pay-per-use just like water, electricity, gas. With the coming of the big data age, in the cloud environment, hosts number has already reached thousands of scale and the cluster size has become more and more large. However, the host nodes are heterogeneous, diverse, uncertain and ambiguous. When faced with high-dimensional matrix computing, traditional fuzzy clustering algorithm is low operational efficiency, lack of space and other operational problems, leading to clustering taking too long to meet the cloud computing requirement. Meanwhile, there are many shortages in the task scheduling based on resource clustering. Most algorithms only consider complete time, not taking account of user’s quality of service and system’s load balancing. Therefore, it’s high time to optimize and improve the traditional serial fuzzy clustering algorithm and task scheduling algorithm, in order to be better applied to cloud computing.In this paper, the problem of resources partition and task scheduling were researched in the cloud computing environment. The main work was as follows:1. Faced with high-dimensional matrix computing, traditional fuzzy clustering algorithm was low operational efficiency and insufficient computation space. Therefore, a fuzzy c-means clustering algorithm was proposed based on Map Reduce parallelization. This algorithm shortened the clustering time and greatly improved the running efficiency.2. Considered the drawbacks of the traditional task scheduling algorithms, this article proposed improvements in the basics of classical Min-Min algorithm, taking into account the user’s Qo S, effectively solving the load imbalance problem caused by traditional Min-Min algorithm and greatly improving user’s satisfaction.3. Expand, recompile and run the Cloud Sim cloud simulation platform to realize the algorithm proposed in this article, which was the resources clustering and task scheduling algorithm based on user’s Qo S. And then compared with other algorithms, the experiment results showed that this algorithm could effectively carried out the task scheduling, ensured the overall task completion time and optimal service cost, and improved user’s satisfaction.