Biometrics in the Cloud: Challenges and Research Opportunities

Privacy is one of several key concerns in the current “cloudification” trend, such as ensuring the privacy of user data and transactions during the outsourcing of data, applications, and infrastructure to the cloud. This is due to a wide range of factors, such as changing societal expectations, and the delegation of the control of the data and transactions to geographically distributed datacenters, including in regimes that may not share the same underpinning legislative framework. For example, European data protection laws may impose additional constraints on the handling and processing of data that are transferred, or restrict the transfer of data, to datacenters located in the US. Designing secure data access and authentication in a cloud computing environment to ensure privacy, particularly in federated or multicloud environments, remains a challenging research agenda.1 One particular research direction is biometric authentication (replacing conventional username and alphanumeric password credentials), and in recent years, Biometrics-as-a-Service (BaaS) (offered over the cloud).2 The latter can be cost-effective, scalable, reliable, hardware agnostic, and provides enhanced security accessible anytime and anywhere (see also Figure 1). As our society becomes increasingly digitalized (e.g., e-government and e-business services), BaaS is an attractive solution for individuals (e.g., the elderly and people with health-related issues such as dementia and Alzheimer's disease) and content and service providers. For the latter, these providers can outsource access and authentication to competent BaaS providers.