Estimating cellular network performance during hurricanes

https://doi.org/10.1016/j.ress.2009.11.003Get rights and content

Abstract

Cellular networks serve a critical role during and immediately after a hurricane, allowing citizens to contact emergency services when land-line communication is lost and serving as a backup communication channel for emergency responders. However, due to their ubiquitous deployment and limited design for extreme loading events, basic network elements, such as cellular towers and antennas are prone to failures during adverse weather conditions such as hurricanes. Accordingly, a systematic and computationally feasible approach is required for assessing and improving the reliability of cellular networks during hurricanes. In this paper we develop a new multi-disciplinary approach to efficiently and accurately assess cellular network reliability during hurricanes. We show how the performance of a cellular network during and immediately after future hurricanes can be estimated based on a combination of hurricane wind field models, structural reliability analysis, Monte Carlo simulation, and cellular network models and simulation tools. We then demonstrate the use of this approach for assessing the improvement in system reliability that can be achieved with discrete topological changes in the system. Our results suggest that adding redundancy, particularly through a mesh topology or through the addition of an optical fiber ring around the perimeter of the system can be an effective way to significantly increase the reliability of some cellular systems during hurricanes.

Introduction

Cellular networks serve a critical role during and immediately after hurricanes. They allow citizens to communicate with emergency services if land-line communication systems have been lost, and they allow families to communicate and organize response plans. Cellular networks have also been used as a backup communication system for emergency personnel when their primary systems fail. For example, during Hurricane Katrina police radio bands failed due to power outages, combined with a lack of fuel for backup generators among other reasons [1]. However, cellular communication systems are not highly reliable during hurricanes. During recent hurricanes, cellular towers have been damaged and backup generators have failed or run out of fuel, rendering the attached communication equipment inoperable [1], [2]. In addition, microwave and fiber–optic links between cellular towers and backbone systems have also failed, rendering that portion of the system equally inoperable [2]. Estimating the reliability of cellular networks during hurricanes is an important problem, and methods are needed that incorporate the effects of failures in multiple nodes and links together with the effects of changes in network traffic during and immediately after a hurricane. At the same time, these methods must utilize appropriate models of the reliability of cellular towers, antennae, and support equipment such as backup generators during hurricanes.

The building blocks of an approach for estimating cellular network reliability are available; yet developing an accurate and efficient estimation model remains a challenge. The first building block includes structural reliability analysis methods for modeling the performance of cellular communication towers under hurricane wind loadings, fault trees, event trees, and other reliability analysis methods for modeling the performance of individual cellular nodes (e.g., a tower, antennae, an external power source, and a backup power source). The second building block includes traffic models for estimating the subscriber traffic and call arrival rate at each cellular tower. The last building block includes simulation techniques for estimating the performance of the cellular system over wide geographical areas. These techniques should accurately simulate the routing and restoration protocols that determine the network availability for given physical states of the system. In this paper we show how this interdisciplinary set of approaches can be combined to develop an integrated approach for assessing cellular network reliability during hurricanes. We demonstrate this approach using a synthetic cellular network.

This paper is organized as follows. Section 2 provides an overview of relevant past work together with the suggested integrated approach. Section 3 presents the case study system that we use to demonstrate the approach. Section 4 discusses the results for this case study area. Section 5 summarizes our approach and its performance.

Section snippets

Cellular network performance and modeling

A typical cellular network is comprised of a set of cellular towers, each of which covers a specific area [3]. A tower's coverage is roughly hexagonal in shape with a radius of about 6.44 km (4 mi). Towers also communicate with each other via two possible methods. The first is the use of buried fiber–optic links between the tower and the local switching station. This method is the most reliable and provides the ability to carry the most calls, but it is also the most expensive. A less expensive

Case study network

We used the synthetic city Mesopolis as our case study for demonstrating the integrated modeling approach [16]. Mesopolis is a synthetic city representing a coastal city with a population of approximately 125,000. It was developed at Texas A&M University to serve as a test bed for developing methods used for assessing infrastructure performance. We extended the basic Mesopolis test bed by designing a cellular communication system based on a standard pattern of hexagonal cellular coverage. As

Results

We simulated the behavior of the cellular network across a series of failures. Each tower may lose its cellular antennae, microwave dishes, fiber–optic links, or any combination therein. The set of these failures across all towers defines a network's state. By utilizing the fragility curves in Fig. 1 and the wind speed model in Fig. 6, we calculated estimates of failure event probabilities for each tower. Finally, from the fault trees in Fig. 2, we defined the state of the network. Running this

Discussion and conclusions

Our results show that the reliability of cellular networks during hurricanes can be evaluated using an interdisciplinary approach that combines structural reliability analysis, probabilistic risk analysis, network traffic modeling, and simulation. The methods provide a basis for both assessing the reliability of an existing network in a computationally feasible manner and for assessing improvements in network performance associated with changes in network topology.

Our results also suggest that

Acknowledgement

This work was supported by Grant ECCS-0725823 from the U.S. National Science Foundation. This work was also supported by Qatar Telecom (Qtel), Doha , Qatar. OPNET Modeler was provided under the OPNET University Programs. This support is gratefully acknowledged, but the work does not necessarily represent the opinion or view of the National Science Foundation.

References (17)

  • Z. Huang et al.

    Hurricane simulation techniques for the evaluation of wind speeds and expected insurance losses

    Journal of Wind Engineering and Industrial Aerodynamics

    (2001)
  • LK Comfort et al.

    Communication, coherence, and collective action: the impact of Hurricane Katrina on communications infrastructure

    Public Works Management and Policy

    (2006)
  • K. Banipal

    Strategic approach to disaster management: lessons learned from Hurricane Katrina

    Disaster Prevention and Management

    (2006)
  • Poole Ian. Cellular Communications Explained. Elsevier/Newnes, 2006. ISBN...
  • Tipper D, Charnsripinyo C, Shin H, Dahlberg T. Survivability Analysis for Mobile Cellular Networks. Proceedings...
  • Varshney U, Snow A, Malloy A. In: Reliability and survivability of wireless and mobile networks. IEEE Computer Society...
  • JR Goodman et al.

    Reliability-based design of wood transmission line structures

    Journal of Structural Engineering

    (1983)
  • Melchers RE. Structural reliability: analysis and prediction. Longman Higher Education...
There are more references available in the full text version of this article.

Cited by (0)

View full text