In the existing cloud computing environment, batch scheduling strategies mainly focus on the management of resources allocation. This paper provides the task scheduling algorithm based on service quality which fully considers priority and scheduling deadline.
IntroductionCloud computing is a new business model, and it is a product of information technology [1]. In cloud computing market, user scale is very large, and the QoS(Quality of Service) requests of each task are not identical. These factors make task schedule in cloud computing very complicated. So it is a technical difficulty to schedule the massive tasks in cloud computing in order to get a better scheduling results with the constrain of budget and deadline.The main batch scheduling algrithms are Min-min [2] and DBCT [3] (under the deadline and budget constrain time optimization algorithm). The idea of Min-min algorithm is that the best resources will be allocated to the minimum task, thus the system throughput rate can be improved and the minimum time of the job can be obtained. While the idea of DBCT is to schedule the task on the resource that can complete the task earliest, also request the cost no more than user's budget and the finish time no more than deadline. With the popularity of cloud computing, some users with less budget may also want to use cloud computing; also there are some tasks with sufficient budget and intensive demand of deadline. For these cases, the number of completed tasks of Min-min and DBCT algorithms will be less. There are also a lot of people put forward other algorithms, but some did not consider the deadline [4], some did not consider the user's budget [5]; they did not balance the relationship between the deadline and the budget.From the perspective of QoS, this paper proposes DBCMNCT (deadline and budget constrained maximized number of completed tasks scheduling algorithm) algorithm in order to meet the budget and deadline constrains. Compared with Min-min and DBCT algorithm, DBCMNCT algorithm can accomplish higher priority tasks better, at the same time the new algorithm can maximize the number of completed tasks. So the user's demands can be meet; on the other hand, the cloud computing can attract more consumers and get more benefits.