Toward scalable and high performance data processing for Cellular Data Network

Cellular Data Network measurement and analysis has long been used to characterize Cellular Data Network usage and user behaviors, but with the rapid growing of all kinds of smart terminals and the varied applications on them, the amount of Cellular Data Network data has been growing exponentially. With respect to this challenge, a new management system for the data storage and query has been emerging. In this paper, we present our applied Hadoop-based system with Cellular Data Network flow upload and analyze components. It combines data uploading and CDR (Calling Detail Records) query with Map/Reduce framework and HDFS. It uses Map/Reduce framework to process the users' request, HDFS to manage the data flow files. Our system has the ability to upload and for the network manager to query the CDR more conveniently and reliable. More importantly, our system is applied to a real-world and has a good scalability, effectiveness and efficiency.