Geoff Kuenning

TrueErase: per-file secure deletion for the storage data path

The ability to securely delete sensitive data from electronic storage is becoming important. However, current per-file deletion solutions tend to be limited to a segment of the operating systems storage data path or specific to particular file systems or storage media. This paper introduces TrueErase, a holistic secure-deletion framework. Through its design, implementation, verification, and evaluation, TrueErase shows that it is possible to build a legacy-compatible full-storage-data-path framework that performs per-file secure deletion and works with common file systems and solid-state storage, while handling common system failures. In addition, this framework can serve as a building block for encryption- and tainting-based secure-deletion systems.

Newer Is Sometimes Better: An Evaluation of NFSv4.1

The popular Network File System (NFS) protocol is 30 years old. The latest version, NFSv4, is more than ten years old but has only recently gained stability and acceptance. NFSv4 is vastly different from its predecessors: it offers a stateful server, strong security, scalability/WAN features, and callbacks, among other things. Yet NFSv4s efficacy and ability to meet its stated design goals had not been thoroughly studied until now. This paper compares NFSv4.1s performance with NFSv3 using a wide range of micro- and macro-benchmarks on a testbed configured to exercise the core protocol features. We (1) tested NFSv4s unique features, such as delegations and statefulness; (2) evaluated performance comprehensively with different numbers of threads and clients, and different network latencies and TCP/IP features; (3) found, fixed, and reported several problems in Linuxs NFSv4.1 implementation, which helped improve performance by up to 11X; and (4) discovered, analyzed, and explained several counter-intuitive results. Depending on the workload, NFSv4.1 was up to 67\% slower than NFSv3 in a low-latency network, but exceeded NFSv3s performance by up to 2.9X in a high-latency environment. Moreover, NFSv4.1 outperformed NFSv3 by up to 172X when delegations were used.

A long-term user-centric analysis of deduplication patterns

Deduplication has become essential in disk-based backup systems, but there have been few long-term studies of backup workloads. Most past studies either were of a small static snapshot or covered only a short period that was not representative of how a backup system evolves over time. For this paper, we collected 21 months of data from a shared user file system; 33 users and over 4,000 snapshots are covered. We analyzed the data set for a variety of essential characteristics. However, our primary focus was individual user data. Despite apparently similar roles and behavior in all of our users, we found significant differences in their deduplication ratios. Moreover, the data that some users share with others had a much higher deduplication ratio than average. We analyze this behavior and make recommendations for future deduplication systems design.

TrueErase: Leveraging an Auxiliary Data Path for Per-File Secure Deletion

One important aspect of privacy is the ability to securely delete sensitive data from electronic storage in such a way that it cannot be recovered; we call this action secure deletion. Short of physically destroying the entire storage medium, existing software secure-deletion solutions tend to be piecemeal at best -- they may only work for one type of storage or file system, may force the user to delete all files instead of selected ones, may require the added complexities of encryption and key storage, may require extensive changes and additions to the computers operating system or storage firmware, and may not handle system crashes gracefully. We present TrueErase, a holistic secure-deletion framework for individual systems that contain sensitive data. Through design, implementation, verification, and evaluation on both a hard drive and NAND flash, TrueErase shows that it is possible to construct a per-file, secure-deletion framework that can accommodate different storage media and legacy file systems, require limited changes to legacy systems, and handle common crash scenarios. TrueErase can serve as a building block by cryptographic systems that securely delete information by erasing encryption keys. The overhead is dependent on spatial locality, number of sensitive files, and workload (computational- or I/O-bound).

vNFS: Maximizing NFS Performance with Compounds and Vectorized I/O

Modern systems use networks extensively, accessing both services and storage across local and remote networks. Latency is a key performance challenge, and packing multiple small operations into fewer large ones is an effective way to amortize that cost, especially after years of significant improvement in bandwidth but not latency. To this end, the NFSv4 protocol supports a compounding feature to combine multiple operations. Yet compounding has been underused since its conception because the synchronous POSIX file-system API issues only one (small) request at a time. We propose vNFS, an NFSv4.1-compliant client that exposes a vectorized high-level API and leverages NFS compound procedures to maximize performance. We designed and implemented vNFS as a user-space RPC library that supports an assortment of bulk operations on multiple files and directories. We found it easy to modify several UNIX utilities, an HTTP/2 server, and Filebench to use vNFS. We evaluated vNFS under a wide range of workloads and network latency conditions, showing that vNFS improves performance even for low-latency networks. On high-latency networks, vNFS can improve performance by as much as two orders of magnitude.

Software engineering education via the use of corporate-sponsored projects: A panel discussion of the approaches, benefits, and challenges for industry-academic collaboration

In this panel, we will address questions regarding the development and execution of one type of industry-academic collaboration - corporate-sponsored projects. As representatives of programs at both public and private institutions, we will address a set of questions related to these programs. We have prepared an initial list that will be augmented with questions submitted and voted on by conference participants in advance of the session. After we have presented our prepared responses, the balance of the session will be dedicated to open discussion with audience members.

Cluster and Single-Node Analysis of Long-Term Deduplication Patterns

Deduplication has become essential in disk-based backup systems, but there have been few long-term studies of backup workloads. Most past studies either were of a small static snapshot or covered only a short period that was not representative of how a backup system evolves over time. For this article, we first collected 21 months of data from a shared user file system; 33 users and over 4,000 snapshots are covered. We then analyzed the dataset, examining a variety of essential characteristics across two dimensions: single-node deduplication and cluster deduplication. For single-node deduplication analysis, our primary focus was individual-user data. Despite apparently similar roles and behavior among all of our users, we found significant differences in their deduplication ratios. Moreover, the data that some users share with others had a much higher deduplication ratio than average. For cluster deduplication analysis, we implemented seven published data-routing algorithms and created a detailed comparison of their performance with respect to deduplication ratio, load distribution, and communication overhead. We found that per-file routing achieves a higher deduplication ratio than routing by super-chunk (multiple consecutive chunks), but it also leads to high data skew (imbalance of space usage across nodes). We also found that large chunking sizes are better for cluster deduplication, as they significantly reduce data-routing overhead, while their negative impact on deduplication ratios is small and acceptable. We draw interesting conclusions from both single-node and cluster deduplication analysis and make recommendations for future deduplication systems design.

Introduction to the Special Issue on USENIX FAST 2017

This special issue of the ACM Transactions on Storage presents some of the highlights of the 15th USENIX Conference on File and Storage Technologies (FAST’17). In the 15 years since its inception, FAST has grown into a damn active community, attracting 116 submissions and over 400 attendees in 2017. FAST’17 continues the tradition of bringing together storage-system researchers and practitioners to explore new directions in the design, implementation, evaluation, and deployment of storage systems. FAST takes a broad view of storage systems, encompassing everything from lowlevel storage devices to information management systems. We selected five high-quality articles for publication in this special issue of ACM Transactions on Storage. The first article, which was also selected as one of the best papers at the conference, is “Application Crash Consistency and Performance with CCFS,” by Thanumalayan Sankaranarayana Pillai, Ramnatthan Alagappan, Lanyue Lu, Vijay Chidambaram, Andrea C. Arpaci-Dusseau, and Remzi H. Arpaci-Dusseau. This article introduces a crash-consistent file system that uses a stream abstraction to guarantee that intra-stream operations are committed in program order to improve the correctness of application crash consistency, while removing ordering constraints across streams to maintain high performance. The second article is “Redundancy Does Not Imply Fault Tolerance: Analysis of Distributed Storage Reactions to File-System Faults,” by Aishwarya Ganesan, Ramnatthan Alagappan, Andrea C. Arpaci-Dusseau, and Remzi H. Arpaci-Dusseau. The authors analyze eight popular distributed storage systems and uncover significant problems that can result in data loss, corruption, and unavailability, with implications for the design of future fault-tolerant distributed systems. The third article is “vNFS: Maximizing NFS Performance with Compounds and Vectorized I/O,” by Ming Chen, Geetika Babu Bangera, Dean Hildebrand, Farhaan Jalia, Geoff Kuenning, Henry Nelson, and Erez Zadok. This article leverages the NFSv4 compounding feature, exposing a highlevel vectorized API to support bulk operations, achieving dramatic speedups on a wide range of workloads under various network latency conditions. The fourth article is “Tiny-Tail Flash: Near-Perfect Elimination of Garbage Collection Tail Latencies in NAND SSDs,” by Shiqin Yan, Huaicheng Li, Mingzhe Hao, Michael Hao Tong, Swaminathan Sundararaman, Andrew A. Chien, and Haryadi S. Gunawi. A well-known problem with flash storage is the issue of long, unpredictable delays induced by garbage collection in the flash translation layer. The authors use a combination of novel strategies to optimize garbage collection, achieving tail latencies that approach the ideal of a system without garbage collection. The final article, also selected as one of the best papers at the conference, is “Efficient Free Space Reclamation in WAFL,” by Ram Kesavan, Rohit Singh, Travis Grusecki, and Yuvraj Patel. The NetApp WAFL file system uses copy-on-write for high-performance writes and efficient snapshots, increasing the demand for free-space reclamation. The authors present the evolution of algorithms and data structures spanning more than a decade of production deployments.