Abstract
Ubiquitous networks enable mobile users to communicate with each other efficiently and independently without the need of inventing agent. This approach is proved to be delay and spectral efficient. Due to the nature of underlying Big data, such networks are prone to several security and privacy challenges. Because such gigantic data is not only difficult to store, maintain and manipulate but Big data's open architecture makes the security threats inevitable. Therefore, incorporating authentication between mobile node and foreign network during roaming in ubiquitous networks has become a tedious task. Very recently, Farash et al. found a number of weaknesses in the schemes proposed by Wen et al. and Shin et al. Furthermore, Farash et al. proposed an enhanced scheme for roaming user in ubiquitous network. However, after thorough investigation, we show fragilities of Farash et al.'s scheme against (1) mobile user anonymity violation; (2) disclosure of secret parameter of mobile node; (3) session key disclosure; and (4) mobile node impersonation attacks. Therefore, we propose an improved scheme to fix these fragilities. We analyze the security of proposed scheme using popular automated verification tool ProVerif. The analysis confirms that the proposed scheme resists the known attacks while having quite low overhead as compared with Farash et al.'s scheme. Therefore, in order to get better performance proposed scheme is a suitable candidate to be employed along with supercomputing systems for dealing the security challenges of big data in ubiquitous networks.