Janki Bhimani

Florida International University

Lecture Information:
  • November 22, 2019
  • 2:00 PM
  • CASE 241

Speaker Bio

Janki Bhimani’s primary research focus revolves around Flash-Based Storage Systems, Big Data Processing, Cloud Computing, High-Performance Computing, and Parallel and Distributed Computing. Her research interest also includes Performance Modeling, Resource Management, and Capacity Planning for various emerging inter-disciplinary research domains. With her extraordinary expertise and extensive experience in the field of new emerging flash-based storage systems and devices, she has made significant contributions to the data storage management community. She is the recipient of the Outstanding Graduate Research Award of 2019 from Northeastern University. She also received Best Paper Awards from flagship conferences. Her work is published in highly selective conferences and journals. She is also the main inventor of top graded patents. Hand-in-hand with her research, she is very passionate about teaching and mentoring. Prior to joining at Florida International University, she previously served Northeastern University as an instructor. She also closely worked with research scientists at Samsung Semiconductor Research Labs towards evolving flash-based SSDs. In her free time, Janki is a creative visual artist. Far from home, amidst nature, she finds her inspiration to paint. She enjoys understanding the impact of art on human psychology, and she can painterly bring motivation, healing, and encouragement through the canvas.


A new era of “Data Age” is approaching today. Data is the fuel for analytics of all the emerging technologies of Internet-of-Things (IoT) and cloud computing. Data management plays a critical role in delivering real-world impacts. However, it is challenging for any systems to efficiently manage data to achieve low latency, high throughput, and good endurance. A good ecosystem is highly demanded to provide fine coordination among multiple facets of the system, such as parallel computing, hierarchical memory caching, and low I/O latency storage. Mitigating bottlenecks at each facet/layer is essential for accelerating overall production-scale deployments. Therefore, in this talk, I will present our research that aims to develop new data management techniques and schemes for building an efficient and reliable computing system for data-intensive applications. Specifically, there are three main components of data management, i.e., data generation, data categorization, and data storage. I will first introduce our primary research focus related to these three data management components. Then, I will mainly elaborate on two research works. First, I will discuss our framework that we are designing to capture enterprise SSD behavior encapsulating various real workloads and system configurations. Second, I will present our new key-value based storage infrastructure for parallel computing using emerging key-value SSDs for accelerating OpenMP high-performance computing (HPC) applications. Finally, I will discuss the challenges in efficient and reliable management of upcoming heterogeneous data-centric computing requirements and possible approaches toward developing infrastructures to accommodate these future requirements.