Under Hadoop cloud platform based hybrid scheduling method of load monitoring

The present invention discloses the next Hadoop cloud platform based hybrid scheduling method of load monitoring. By analyzing the effect of Max-D scheduling algorithm, FIFO algorithm and fair scheduling algorithms and their application scenarios, proposed a strategy based hybrid scheduling load monitoring. The present invention, by real-time monitoring of system load, choose the most appropriate current load conditions scheduling algorithm three algorithms. Compared to a single use of a scheduling algorithm based on hybrid scheduling load monitoring it has obvious advantages, able to adapt to changes in Hadoop system loads, enabling the system to maintain a relatively good performance.