Information flow tracking was proposed more than 40 years ago to address the limitations of access control mechanisms to guarantee the confidentiality and integrity of information flowing within a system, but has not yet been widely applied in practice for security solutions. Here, we survey and systematize literature on dynamic information flow tracking (DIFT) to discover challenges and opportunities to make it practical and effective for security solutions. We focus on common knowledge in the literature and lingering research gaps from two dimensions— (i) the layer of abstraction where DIFT is implemented (software, software/hardware, or hardware) and (ii) the security goal (confidentiality and/or integrity). We observe that two major limitations hinder the practical application of DIFT for on-the-fly security applications: (i) high implementation overhead and (ii) incomplete information flow tracking (low accuracy). We posit, after review of the literature, that addressing these major impedances via hardware parallelism can potentially unleash DIFT’s great potential for systems security, as it can allow security policies to be implemented in a built-in and standardized fashion. Furthermore, we provide recommendations for the next generation of practical and efficient DIFT systems with an eye towards hardware-supported implementations.