Opportunities and challenges of wireless networks in the era of mobile big data

Jin Xu, Qimei Cui, Yanzhao Hou, Ping ZhangChinese 2015 · DOI

Abstract

With the rapid development of information and communication technologies such as mobile Internet, cloud computing, and the Internet of Things, the information society has stepped into the networked big data era. Smart terminals, widely applied in everyday life, are driving the explosive demands of wireless data services. In the era of mobile big data, wireless networks face with many challenges owing to the specific characteristics of big data such as high volume, variety, and non-uniform distribution in the time and space domains. And new approaches are required to address these new challenges. On the one hand, new spectrum technology, transmission technology, and intelligent networking offer promising evolution directions. On the other hand, mining the information contained in the low-density mobile big data, could be a new way to optimize current wireless networks. Data mining theory incorporated with mobile Internet can be applied in a wireless network to realize flexible networking, optimization of radio resource configuration, and green communication.

Introduction

The "Mobile Big Data" (MBD) era is fundamentally reshaping wireless communication. While the explosion of data poses severe challenges to traditional network architectures, the data itself contains the "hidden treasure" needed to build more intelligent, flexible, and green 5G networks. Transitioning from "network-centric" to "data-centric" architectures allows engineers to use data mining to predict user behavior, optimize resource allocation, and significantly reduce energy consumption.

The Challenges of Mobile Big Data

Traditional "one-size-fits-all" network designs are struggling with the four "Vs" of Big Data:

  1. Volume: Global mobile traffic is growing exponentially, reaching 1.5 EB/month in 2013.
  2. Variety: Traffic has shifted from simple voice/SMS to complex, bursty applications like high-definition video and IoT data.
  3. Velocity: Data generation and processing must occur at lightning speeds.
  4. Value: While raw data has low "value density," the insights buried within are vital for optimization.

Current networks are often static and designed for "worst-case" scenarios, leading to massive energy waste. This is particularly inefficient given that traffic is highly non-uniform across both geography and time.

Spatio-temporal non-uniform distribution of mobile traffic in urban environments.

As shown in the spatio-temporal traffic maps, demand is concentrated in specific "hotspots" and varies significantly over time. Approximately 70% of traffic typically originates from these localized areas, suggesting that "Ultra-Dense Networks" (UDNs) and "Small Cells" are essential for future capacity.

Data-Driven Wireless Networks

To address these challenges, the paper proposes a framework for Data-Driven Wireless Networks. Key methodologies include:

  • Dynamic Network Deployment: Utilizing C-RAN (Cloud Radio Access Network) and SDN/NFV to allow the network to "follow the data" by dynamically reallocating resources.
  • Predictive Analytics: Using algorithms like K-means clustering to group users by behavior and predict mobility patterns.
  • QoE-Oriented Management: Shifting optimization targets from technical Quality of Service (QoS) metrics to user-perceived Quality of Experience (QoE).

Green Communication and Energy Efficiency

One of the most critical applications of MBD is in "Green Communication." Base stations are the primary drivers of network operational costs and energy use.

Energy consumption breakdown of wireless network components.

As illustrated in the energy breakdown, base stations consume 57% of a network's total energy. By using MBD to predict traffic lulls, specific base stations can be put into "sleep mode" without affecting user experience, significantly reducing the environmental and financial footprint of the network.

Conclusion and Future Directions

While MBD offers immense opportunities, several hurdles remain. The tension between data utility and personal privacy requires new "middle ground" solutions. Furthermore, the computational complexity of real-time big data processing for millisecond-level radio resource management is significant. Future work must bridge the gap between high-level data mining theories and the low-level physics of wireless signal propagation to create truly self-organizing, intelligent networks.

Reference: Zhang P, Cui Q M, Hou Y Z, et al. Opportunities and challenges of wireless networks in the era of mobile big data. Chin Sci Bull, 2015, 60: 433–438.

Summary updated

Abstract

With the rapid development of information and communication technologies such as mobile Internet, cloud computing, and the Internet of Things, the information society has stepped into the networked big data era. Smart terminals, widely applied in everyday life, are driving the explosive demands of wireless data services. In the era of mobile big data, wireless networks face with many challenges owing to the specific characteristics of big data such as high volume, variety, and non-uniform distribution in the time and space domains. And new approaches are required to address these new challenges. On the one hand, new spectrum technology, transmission technology, and intelligent networking offer promising evolution directions. On the other hand, mining the information contained in the low-density mobile big data, could be a new way to optimize current wireless networks. Data mining theory incorporated with mobile Internet can be applied in a wireless network to realize flexible networking, optimization of radio resource configuration, and green communication.
BibTeX
@inproceedings{xu2015opportunities,
  author = {Jin Xu and Qimei Cui and Yanzhao Hou and Ping Zhang},
  title = {{Opportunities and challenges of wireless networks in the era of mobile big data}},
  booktitle = {Chinese},
  year = {2015},
  doi = {10.1360/n972014-01165},
}
Figure