Joy Arulraj - Research Statement

Modern data-intensive applications apply statistical analysis algorithms onmassive databases to deliver qualitatively better results in many domains, including science, governance, and business. These applications require low latency, always-on, and cost-effective data management. This places increased demands on today’s database management systems (DBMSs) whose architectures are tailored for the hardware landscape of the s. The capabilities of nascent hardware technologies invalidate the long-held design assumptions these systems make about memory and storage. This shift argues for research into rearchitecting DBMSs for a new era of heterogeneous, dis-aggregated, and domainspeci c hardware architectures. Given this outlook, the central theme of my research is on the development of new database management systems that leverage the characteristics of emergent hardware technologies to meet the requirements of data-intensive applications. In particular, my current research focuses on a new class of non-volatile memory (NVM) technologies that blur the gap between volatile memory and durable storage. NVM supports low latency byte-addressable accesses similar to DRAM, but all writes are persistent like SSDs. There are several aspects of NVM that make existing DBMS architectures inappropriate for them. My research investigates how to rearchitect the DBMS from the ground-up to take advantage of NVM []. I redesign the fundamental algorithms and data structures employed in traditional DBMSs to leverage the persistence and performance characteristics of NVM. This enables the DBMS to support low latency transactions, instantaneous recovery from system failures, and cost-effective data management. My work shows that NVM’s impact straddles across all the layers of the DBMS, including logging and recovery [], storagemanagement [, ], indexing [], and query execution []. I plan to continue collaborating with partners in both academia and industry to identify shortcomings in existing systems for supporting data-intensive applications. I will take a principled approach to solving these problems by devising analytical models informed by both theory and practice. I will then validate these models by building systems and evaluate their efficacy using scienti c experimentation. With the widespread adoption of systems for performing large-scale data science, my research will help move the state-of-the-art forward in this eld. I will next discuss the two primary areas of my current research agenda: non-volatile memory and self-driving DBMSs. These projects are a testament to my overall vision of redesigning DBMSs to take advantage of breakthroughs in other branches of computer science. I conclude with a discussion of the research directions that I intend to pursue in future.