Projects

How FIFO and CLOCK respond to frequency skew

We extend van den Berg and Toswley’s proof of the Schur-convexity of FIFO miss ratio to explain why FIFO out-performs RANDOM on real workloads, while using our proofs to explain why CLOCK often out-performs LRU.

QoS, Fairness and Isolation in One Formula

A unified linear programming formula for optimal per-VM IOPS allocation under QoS and device constraints.

A Distributed Object Store built on Conventional File System

This is an emulation of a distributed store system built on Zoned Namespace SSDs. The system guarantees strong consistency and fault tolerant.

Emulating an ISC-enabled LSM-tree for Read Optimization

This project presents a proof of concept for an ISC-enabled LSM-tree, aiming to enhance read performance and space efficiency by offloading parallel search computations from the host to storage devices.

Configurable and Cache-Accurate Trace Generation for Storage Benchmarking

2DIO is a configurable I/O trace generator capable of producing cache-accurate workloads with tunable performance cliffs and plateaus.

A library for cache simulation

This package implements a novel algorithm for the LRU cache eviction policy using a fixed-size array.

Generating high-fidelity I/O traces with GAN

Reproduction of LSTM-based GAN synthetic I/O generation proposed by Zhang et al. (2024).