Abstract: The query processing and optimization of data stream is becoming a hotspot in the International database.The arrival rate of tuples in a data stream can be unpredictable and bursty.These irregular changes in the input rate can result in dramatic fluctuation of the system load over time.When input rates exceed the system processing capacity，the system will become overloaded and the performance will deteriorate.To solve this problem，load shedding is one of promising approaches.In this paper，some key techniques are discussed，such as，the prediction of system processing load，the optimal placement of a load shedder in a query plan，the combination of load shedder and how much load to shed.

Key words: load shedding, data stream, query processing

摘要： 流数据连续查询及优化研究已成为当前国际数据库研究领域的一个热点。数据流的到达速率经常是不可预测的且具有很高的突发性，数据流速这种不规则的变化会引起系统负载急剧的波动。当输入速率超过系统处理能力时，系统会发生过载并且导致系统性能的恶化，降载技术是解决此问题最有效的途径之一。对降载技术中系统负载估计、降载器的最佳的放置位置、降载量的大小、降载器合并等关键问题进行了讨论。

关键词: 降载, 数据流, 查询处理