在基于Web的数据库应用中,读取和传输含有大对象(LOB)类型数据的查询结果集时若采用传统的方法易导致服务器端内存不足和网络拥塞。提出一种数据结果集块状读取(SBBR)算法,可以基于动态设定的内存块大小来读取结果集。该算法已应用到科学数据网格的数据访问中间件DAS系统中。测试结果表明,SBBR算法能较好地处理包含LOB数据的结果集。
In the application of database on Web,when Large OBject(LOB) type columns exist in the result set of a database querying,it can easily result in some problems like running out of memory of server and network traffic if using the general methods to read and transfer the result set in the network. A novel algorithm——SBBR is provided to avoid those problems,it can handle the result set in chunks and the block size can be dynamically set. SBBR is implemented in the grid middleware-DAS of science data grid. Test results show it can handle the result set which includes LOB well.