Detection of Distributed Denial Of Service (DDoS) attacks have been increasing rapidly in the recent years mainly causing a severe risk to security of network systems in large and also occupying resources of computer systems such as computer power, memory, band width and many more. Detection and Prevention of DDoS attacks at fasters not only helps in the network security but also increases computer system performance by not occupying computer resources. Therefore, it has become very essential in today’s fast growing world to detect and prevent DDoS attacks. The proposed system makes use of Hadoop Framework which includes the techniques of map reduce that uses the log files as the datasets dividing them into smaller multiple chunks for easy processing and training the proposed model for detecting and preventing the network systems from DDoS attacks at faster manner..