Florida International University

Janki Bhimani

Raju Rangaswami

Adnan Maruf

Ashikee Ghosh

Manoj Pravakar Saha

Systems and methods for optimizing the internal metadata management of key-value solid state drives (KVSSDs) are provided. A key-to-physical-address (K2P) mapping scheme (resizable hash-based indexing) that focuses on index resizing and space optimization can be used. The resizable metadata management scheme can efficiently handle a highly variable workload in an incredibly large keyspace. It can provide predictable metadata access cost, as well as fast membership checking. The metadata management structure can be designed to ensure near-constant performance in terms of metadata access cost and fast membership checking.

Systems and methods for optimizing data management within key value storage

Systems and methods for dynamically placing data in a hybrid memory structure are provided. A machine learning (ML)-based, adaptive tiered memory system can actively monitor application memory to dynamically place the right data in the right memory tier at the right time. The memory system can use reinforcement learning to perform dynamic tier placement of memory pages.

Growing community of inventors

Miami, FL, United States of America

Janki Bhimani