Analyzing the Security Susceptibilities of Biometrics Integrated with Cloud Computing

Analyzing the Security Susceptibilities of Biometrics Integrated with Cloud Computing

John R. Regola (Pennsylvania State University – Altoona, PA), John K. Mitchell III (Pennsylvania State University – Altoona, PA), Brandon R. Baez (Pennsylvania State University – Altoona, PA) and Syed S. Rizvi (Pennsylvania State University – Altoona, PA)
DOI: 10.4018/978-1-4666-6559-0.ch009

Abstract

In the present scenario, the vulnerabilities associated with cloud computing and biometric technology rank among the most vital issues in information security. In this chapter, the primary goal is to investigate the physical and informational security susceptibilities of biometrics, analyze the structure and design possibilities of the cloud, and examine the new developments of biometrics with cloud computing. Foremost, the authors analyze the developments of biometrics and compare the performance based on defining characteristics. In addition, they examine threats and attacks that can compromise the assets of an organization or an individual's sensitive information. Furthermore, this chapter provides a comprehensive discussion on the physical vulnerabilities of biometrics. Moreover, one section of this chapter focuses on the informational and database vulnerabilities. In this chapter, the authors also discuss the design considerations and cloud computing paradigm in relation to biometric security systems.
Chapter Preview
Top

Introduction

The development of biometric technology has improved exponentially within the last few decades (Sabena, Dehghantanha, & Seddon, 2010). The ever-evolving advancements in this field address security issues that, for many years, security professionals have been trying to remedy. There are two basic classifications of biometrics: physiological and behavioral. In the case of physiological, scanners measure size, shape, and uniqueness of physical characteristics of the human body (More, Ubale, & Jondhale, 2008). The verification of an individual’s identity using physiological biometrics is solely reliant on input that cannot be altered by the users. For instance, fingerprint scanners identify the variations between the ridges, loops, and whirls within a fingerprint, which is unique to every individual. An iris scanner identifies the intricate structures within an iris, patterns that no two humans share. In contrast, behavioral biometrics is the classification that involves monitoring the consistency of behavior in an individual’s input, including, but not limited to, gait, voice recognition, and palm pressure scanning.

The results in these identification entries are subject to variation, and as a result often face issues such as FAR and FRR. FAR refers to the instances in which the scanner falsely grants access to a biometric system. On the contrary, the FRR refers to instances in which the scanner falsely rejects access to a biometric system. Since lowering the FAR would increase the FRR, they need to be balanced in what is known as the equal error rate (EER) (Kukula, Sutton, & Elliott, 2010). The EER is the rate of equality between the FAR and FRR (Kukula, Sutton, & Elliott, 2010).

In order to reduce the FAR and FRR, while maintaining a high level of accuracy, multimodal biometrics can be applied. There are two major types of multimodal systems: intra-class and inter-class. This refers to the use of either multiple layers of the same technology or integrating multiple forms of biometric systems (Chin Ong, Goh, & Hiew, 2009). For example, a multimodal security system can comprise of either different types of fingerprint scanners (such as optical and thermal scanners) or a fingerprint scan followed by an iris scan. Later, in the development of biometric technology section, we will provide several examples of how multimodal biometrics can be implemented.

Currently, there are at least three main characteristics shared among many of the common biometric security systems that we previously discussed. The robustness, intrusiveness, and distinctiveness are qualities that relate to identifying an individual (Govinda & Ngabirano, 2012). Robustness refers to a trait that is subject to change over a period of time. Meanwhile, intrusiveness revolves around the comfort of the individual and their personal privacy when using the biometric system. A less intrusive biometric security system is ideal for acceptance by society. Distinctiveness relates to the variations in a specific trait contained by the general population. The greater the level of distinctiveness, the more exclusive the system would be. These characteristics enable us to compare and contrast them in an effort to understand the advantages and disadvantages of different biometric security systems.

Fingerprint scanning, is fairly robust and highly distinctive, especially when compared to dynamic signature verification. However, there is still a chance that some individuals have unusable fingerprints due to aging or an injury. In general, a conclusion can be made that pattern or geometric scanning is more robust than dynamic scanning. This is evident by the low level of robustness seen in dynamic signature verification and keystroke dynamics (Govinda & Ngabirano, 2012). Another factor to consider is the level of intrusiveness, since fingerprints only require a touch rather than a retinal or facial scan. These require static positioning and keen observation.

Key Terms in this Chapter

Interoperability: Interoperability relates to the interaction between systems and their ability to exchange and use information. For instance, multiple software applications used on two different physical machines from a different vendor should be able to receive and understand each other’s data in an efficient manner.

Hypervisor: A hypervisor is a software program that allows multiple users to access a single hardware host through the use of virtual machines. These virtual machines or guest machines appear to contain all of the resources of the host machine with the hypervisor allocating the resources to each virtual machine based on demand.

Bare-Metal: Bare-metal or Type 1 virtualization is determined by the hypervisor having direct control over the hardware of the host.

Reliability: Reliability is the consistency of a system’s performance according to a predefined set of conditions. The probability of failure is a deciding factor in determining the reliability of a system. In the context of biometrics, when evaluating the performance, the FAR, FRR, EER, ROC, CMC, and D’ must be considered.

Biometrics: Biometrics refers to a system of measuring physiological or behavioral characteristics in order to verify an individual’s identity. Currently, there are multiple existing biometric systems such as fingerprint scanning, iris recognition, signature verification, etc. which are being implemented by private and government institutions.

Hosted: Hosted or Type 2 virtualization involves the hypervisor running at a second software layer within an operating system environment. Through the host operating system, the hypervisor is able to create guest virtual machines.

Failover: In the event of system failure from a hardware or network malfunction, redundant resources are used to restore the system back to its original state. There is a direct relationship between failover and reliability.

Scalability: Scalability is the ability of a system to adapt to its own expansion while displaying increased efficiency. In other words, if additional network resources are added to a system, a higher level of productivity can be expected. If this is untrue, then the system is not considered scalable.

Node: A node is a location in a network where data is processed. Typically, a node is a physical device that is connected to a network and capable of sending or receiving data. In cloud computing, the concept of a node extends into virtualization where virtual nodes can be leveraged to perform parallel computations.

API: (Application Programming Interface) is the combination of protocols, tools, and routines that enable an easier method for designing and developing software applications. A graphical user interface (GUI) requires these components and therefore may be dependent on the API.

Complete Chapter List

Search this Book:
Reset