Application research on MapReduce programming model in network-I/O-intensive programs

This paper researched whether MapReduce was suitable for slight data sets,network-I/O-intensive programs through the open source project—Hadoop.A typical slight data set,network-I/O-intensive program—the FTP site scanning program was rewritten based on MapReduce;set up a small Hadoop cluster,with the default configurations adjusted,then the performance of the original program and the rewritten one was test respectively on real data.The experiments show that MapReduce programming model is highly distributed,and it's also suitable for slight data sets,network-I/O-intensive programs.