Home Page of Gopala Suryanarayana

Projects @ FSL

Tracefs

Tracefs is a thin stackable file system for capturing file system traces in a portable manner. It can capture uniform traces for any file system, without modifying the file systems being traced. Tracefs can capture traces at various degrees of granularity: by users, groups, processes, files and file names, file operations, and more; it can transform trace data into aggregate counters, compressed, checksummed, encrypted, or anonymized streams; and it can buffer and direct the resulting data to various destinations (e.g., sockets, disks, etc.). The modular and extensible design allows for uses beyond traditional file system traces: Tracefs can wrap around other file systems for debugging as well as for feeding user activity data into an Intrusion Detection System.

Replayfs

Replaying traces is a time-honored method for benchmarking, stress-testing, and debugging systems-and more recently-forensic analysis. One benefit to replaying traces is the reproducibility of the exact set of operations that were captured during a specific workload. Existing trace capture and replay systems operate at different levels: network packets, disk device drivers, network file systems, or system calls. System call replayers miss memory-mapped operations and cannot replay I/O-intensive workloads at original speeds. Replayfs is the first system for replaying file system traces at the VFS level. The VFS is the most appropriate level for replaying file system traces because all operations are reproduced in a manner that is most relevant to file-system developers

A Low-Complexity Versatile Transactional File Interface

A transactional file system has long been desired, and already some systems have been proposed. Each time, however, the proposal focuses on augmenting a traditional file system or creating an entirely new one, usually based upon a larger and more complicated database. Both approaches lead to kernel bloat: the operating system becomes increasingly large and complicated, incorporating a more powerful abstraction at the expense of performance, maintainability, and simplicity. Fortunately, there is an easier way. We show that there is a small set of kernel modifications that in concert with a carefully constructed user library can enable transactional file system operations. We demonstrate that with these modifications, one can construct a small, portable user-level library that offers transactional semantics as powerful and efficient as previous in-kernel approaches, with no overhead on non-transactional applications.

Ext3 Modifications To Support Transactional Semantics

Motivated by the powerful atomicity guarantees that the ext3 offers today for smaller granularity operations, we are working on the extension of this onto process level transactions. This would like the most efficient way to guarantee ACID semantics to a process while not sacrificing any of the efficiency that the ext3 already offers today.

Internship

VMware (Summer '07)

Para Virtualized File system - A project aimed towards easing management of Virtual Machines.

Gopala Suryanarayana

Welcome

Contact Information