WWW应用的广泛普及使得电子商务快速发展,然而电子商务网站发展的瓶颈也日趋严重,即无论管理人员的经验多么丰富,他们都难以准确预测下一个时间段的访问量。本文详细介绍了服务器工作负载的三种预测技术,并应用这些技术对内蒙古大学学生机房服务器的访问量进行了预测,希望以此来有效避免系统崩溃的可能,并且改善服务器的性能. However, the bottleneck of the development of e-commerce websites is also aggravating, that is, no matter how rich the experience of managers, it is difficult for them to accurately predict the traffic of the next time period. This article details three forecasting techniques for server workloads, and uses these techniques to predict the number of student room servers in Inner Mongolia University. We hope this will effectively prevent the possibility of system crashes and improve server performance.