Encyclopedia of Wireless Networks

Living Edition
| Editors: Xuemin (Sherman) Shen, Xiaodong Lin, Kuan Zhang

Wireless Big Data

Living reference work entry
DOI: https://doi.org/10.1007/978-3-319-32903-1_97-1



Wireless big data describe massive data (e.g., wireless signaling data, device logs, browser history, and payment records) generated by wireless networks (e.g., cellular networks, wireless sensor networks, and wireless ad hoc networks), as well as related technologies for data collection and analytics.

Historical Background

Similar to the traditional concept of big data, wireless big data can be also described by four “V”: volume, variety, velocity, and value. In addition, wireless big data show unique characteristics of large-scale distribution and strong heterogeneity, which impose additional challenges in data collection and analytics.

Wireless big data are generated everywhere in wireless networks, and their collection incurs high communication cost. In the early period, a natural approach for reducing communication cost is to sample a subset of data. In particular, compressive sampling theory (Donoho 2006) has been successfully applied in wireless sensor networks (Luo et al. 2009). It can extract global information by collecting data from a subset of sensors. Later, thanks to the significant improvement of network capacity, it becomes affordable to collect all data to avoid information missing. However, heavy network traffic generated in wireless big data collection still exists, and it can affect other applications running over wireless networks. For example, some network service providers collect system logs generated by cellular network infrastructures (e.g., base stations and switches) for network diagnosis. These traffics occupy a portion of network bandwidth, and they should be well scheduled to minimize the influence to mobile user traffic.

Meanwhile, wireless big data show heterogeneity because they are generated by various devices and applications. They have different formats and are collected with different frequencies. To provide a unified data representation, a tensor model has been proposed to represent the unstructured, semistructured, and structured data generated in wireless networks (Kuang et al. 2014). He et al. (2016) have introduced a unified data model based on the random matrix theory and machine learning.

Wireless big data can be analyzed by techniques (e.g., data mining and pattern recognition) that are also adopted by traditional big data analytics. In addition, some wireless big data can be expressed in spatial-temporal dimensions that impose additional challenges. Xu et al. (2016) design a time series analysis approach that can decompose large-scale mobile traffic into regularity and randomness components. Then, the traffic patterns can be predicted based on regularity components. In recent years, there is a growing trend of applying machine learning techniques, especially deep learning, in wireless big data analytics. Alsheikh et al. (2016) present an overview and brief tutorial on deep learning in mobile big data analytics, as well as a scalable learning framework over Apache Spark. Yang (2016) investigates various learning methodologies for wireless big data analytics. Qian et al. (2017) survey recent research results of wireless big data analytics.

Key Applications

Wireless big data can be applied in many scenarios, such as user mobility modeling, traffic analysis, network planning, privacy, and security enhancement. Some key applications are listed as follows.
  • User Mobility Analysis: Wireless big data generated by mobile phones have great potential for human mobility analysis. By mining wireless big data, the human mobility can be understood, predicted, or even controlled. For instance, by recording and analyzing the time spent at different locations for different individuals, it can be robustly identified whether an individual is at home or at office, which can improve the efficiency urban cellular network planning. Moreover, individual mobility and mobile traffic are closely linked to the social ecology, which helps achieve a better understanding of human behavior.

  • Social Networks: Wireless big data also exist in various mobile social networks. Massive high dimensional personal data are collected from mobile users by the cloud. Analyzing these data can predict individual activities and provide personalized recommendation. A typical example is video recommendation. The user context data can be analyzed by a recommendation model, so that appropriate videos can be recommended for users. Meanwhile, recommendation policy can be improved according to user feedbacks.

  • Security and Privacy: Although wireless big data can help understand individual behavior more easily, the privacy and security issues are inevitable. Wireless big data containing various privacy information, such as user locations, interests, and habits, which would be easily obtained by malicious users due to the broadcast nature of wireless transmission. Therefore, many applications concentrate on protecting privacy and security of wireless big data. For instance, using the method of distorting time, the interest and mobility habits can be concealed for achieve anonymous mobile data set.


  1. Alsheikh MA, Niyato D, Lin S, Tan HP, Han Z (2016) Mobile big data analytics using deep learning and apache spark. IEEE Netw 30(3):22–29.  https://doi.org/10.1109/MNET.2016.7474340 CrossRefGoogle Scholar
  2. Aron M, Brown B (2001) The politics of nature. In: Smith J (ed) The rise of modern genomics, 3rd edn. Wiley, New York, pp 234–295Google Scholar
  3. Donoho DL (2006) Compressed sensing. IEEE Trans Inf Theory 52(4):1289–1306.  https://doi.org/10.1109/TIT.2006.871582 MathSciNetCrossRefzbMATHGoogle Scholar
  4. He Y, Yu FR, Zhao N, Yin H, Yao H, Qiu RC (2016) Big data analytics in mobile cellular networks. IEEE Access 4:1985–1996. https://doi.org/10.1109/ACCESS.2016.2540520 CrossRefGoogle Scholar
  5. Kuang L, Hao F, Yang LT, Lin M, Luo C, Min G (2014) A tensor-based approach for big data representation and dimensionality reduction. IEEE Trans Emerg Top Comput 2(3):280–291.  https://doi.org/10.1109/TETC.2014.2330516 CrossRefGoogle Scholar
  6. Luo C, Wu F, Sun J, Chen CW (2009) Compressive data gathering for large-scale wireless sensor networks. In: Proceedings of the 15th annual international conference on mobile computing and networking. ACM, New York, MobiCom’ 09, pp 145–156.  https://doi.org/10.1145/1614320.1614337 CrossRefGoogle Scholar
  7. Qian L, Zhu J, Zhang S (2017) Survey of wireless big data. J Commun Inf Netw 2(1):1–18.  https://doi.org/10.1007/s41650-017-0001-2 CrossRefGoogle Scholar
  8. Xu F, Lin Y, Huang J, Wu D, Shi H, Song J, Li Y (2016) Big data driven mobile traffic understanding and forecasting: a time series approach. IEEE Trans Serv Comput 9(5):796–805.  https://doi.org/10.1109/TSC.2016.2599878 CrossRefGoogle Scholar
  9. Yang C (2016) Learning methodologies for wireless big data networks. Neurocomput 174(PA):431–438. https://doi.org/10.1016/j.neucom.2015.04.111 CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.School of Computer Science and EngineeringThe University of AizuAizu-WakamatsuJapan

Section editors and affiliations

  • Song Guo
    • 1
  1. 1.Department of ComputingThe Hong Kong Polytechnic UniversityKowloonHong Kong