The Internet has opened up a whole new set of opportunities for advancing the science of psychometrics and the technology of testing. It has also created some new challenges for those of us involved in test design and testing. In particular, we are seeing impacts from internationalization of testing and new models for test delivery. These are changing the traditional balance of power between test producers, test users, test-takers, and the consumers of the results of testing. The use of a relatively new and immature technology for both low and high stakes assessment poses a number of challenges that need to be addressed. This article endeavors to clarify the role and function of test administration. It describes 4 modes of test administration and discusses the implications each has for the degree of control that can be exercised over testing. In this context, it considers issues such as security and confidentiality, including data protection and copyright; technical performance and bandwidth constraints and their impact on testing; and what constitutes "good practice" in the remote delivery of testing.