Tunable Communication for Distributed Systems
(Spring 2025 - Present)
At Brown, I am building a communication layer for distributed
systems. This system has two main contributions:
1) By fitting the guarantees provided by the communication layer to
the exact needs of the distributed system, we are able to improve
the performance of the system.
2) Our abstraction allows a developer to cleanly and rapidly specify
the communication pattern they want for their distributed system.
My distributed communication abstraction, Tundra, utilizes DPDK for
low-latency networking, and has the ability to switch this backend
to kernel sockets or a virtualized debugging interface.
KLean: Extending Operating System Kernels with Lean
(Fall 2025 - Present) At Brown, I am involved in the KLean project. We are attempting to address the limitations of eBPF (limited loop invariants, concurrency, and memory access) by creating a framework for writing kernel extensions in Lean. We presented our idea at PLOS 2025 (a workshop a part of SOSP 2025). If you want to read more about this project, see [ACM Publication] or [KLean PDF]
RDMA Data Structures
(Fall 2023 - Summer 2024) In my undergraduate, I built concurrent data structures using one-sided RDMA. In total, I implemented three novel designs of a B+Tree, skiplist, and an IHT (Inter-locked hashtable). I also designed a software-cache for one-sided RDMA that enabled a high-performance design by optimizing the read-heavy traversals of the tree-like indexes in the B+Tree, skiplist, and IHT.
