Creative malfunction : finding fault with Rowhammer

This essay offers a close reading of a notable computer security vulnerability: the Rowhammer bug in Dynamic Random Access Memory. The story of Rowhammer provides a vivid demonstration of how previously unanticipated possibilities of malfunction emerge as vulnerabilities and, in demanding repair, exert strong pressures on the future development of technologies. Vulnerabilities like Rowhammer, I show, do not appear fully formed; Rowhammer emerged over several years, as successive studies teased out the nature of the problem and invented new methods that put it to use in practical exploits, demonstrating how it can be used to compromise the security of affected systems. These studies cast new light on a variety of existing components, rendered faulty in relation to their failures to contain the error, or in their usefulness for the crafting of an exploit. There is no simple fix for a problem like Rowhammer. Any resolution depends upon a characterisation of the fault, which as we will see can still be subject to further revision years later. I close with an examination of some of the theoretical implications. The study of computer vulnerability, I argue, gives us insights into the historicity of technology and the permanent conditions of change and revision that characterise contemporary computing. Vulnerability research can be understood as a process of real time exploration of computational systems’ ‘adjacent possible,’ creating new ways in which things can be at fault. Drawing on critical studies of repair, I argue that the interventions that arise from these explorations should be understood, not in terms of restoring a system to a prior good state, but as a creative, future-making force