Building a new NVMe Scale-out Distributed Storage Systems from ground up. We use all kinds of latest amazing technologies that fetch you I/O latencies of only a few microseconds.

Non-Profit initiative started by a group of friends to address carbon emission reduction with specific focus on Food Sustainability

Filesystems and Object Storage (StorageGrid)Building a cloud gateway that supports NAS protocols where files can be ingested to Object Storage as SMB or NFS files.C/C++, Go, Scala, Python

Deduplication service for a NFS filesystemC, Python

Research areas:* (C++ and Python) Workload analytics to determine workload patterns and characteristics and use them to improve system performance. Includes, infrastructure performance analytics and modeling, adaptive Flash Read Cache for shifting working sets, predicting application characteristics from block traces through data analytics for self-adaptive storage system without prior knowledge about application, helping customers understand their workloads with respect to I/O patterns, working set profiles, data locality, extent hotness and cache/volume/drive/CPU statistics collected in storage system and visualize them.* (C++) Volume backup to cloud object storage : Volume backup to cloud object storage with high degree of data synchronization and atomicity between local copy and cloud, without compromising on access latency and data correctness.* (Golang) Persistent Storage Plugin for Kubernetes - Open source Flex Volume Driver provided by Kubernetes for Data Storage vendors.* Online data migration between active storage arrays.* InterVM Communication Channel - Virtualized storage platform.* Differentiated Storage Architecture - A workload-aware tiered storage

* File System I/O Access Hints for Performance Enhancement at Block Layer* Object Pre-fetching from Cloud Storage* Caching And Pre-fetching Techniques in Data De-duplication * Methods to Improve Read Performance in Data De-duplicationAdvisor : Prof. David H.C Du

EMC Certified Information Storage Associate