SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
BurstFS is a distributed file system for node-local burst buffers on high performance computing systems. BurstFS presents a shared file system space across the burst buffers so that applications that use shared files can access the highly-scalable burst buffers without changing their applications.
The Oak Ridge Leadership Computing Facility (OLCF) runs the No. 4 supercomputer in the world, supported by a petascale file system, to facilitate scientific discovery. In this paper, using the daily file system metadata snapshots collected over 500 days, we have studied the behavioral trends of 1, 362 active users and 380 projects across 35 science domains. In particular, we have analyzed both individual and collective behavior of users and projects, highlighting needs from individual communities and the overall requirements to operate the file system. We have analyzed the metadata across three dimensions, namely (i) the projects' file generation and usage trends, using quantitative file system-centric metrics, (ii) scientific user behavior on the file system, and (iii) the data sharing trends of users and projects. To the best of our knowledge, our work is the first of its kind to provide comprehensive insights on user behavior from multiple science domains through metadata analysis of a large-scale shared file system. We envision that this OLCF case study will provide valuable insights for the design, operation, and management of storage systems at scale, and also encourage other HPC centers to undertake similar such efforts.
The Oak Ridge Leadership Computing Facility (OLCF) has deployed multiple large-scale parallel file systems (PFS) to support its operations. During this process, OLCF acquired significant expertise in large-scale storage system design, file system software development, technology evaluation, benchmarking, procurement, deployment, and operational practices. Based on the lessons learned from each new PFS deployment, OLCF improved its operating procedures, and strategies. This paper provides an account of our experience and lessons learned in acquiring, deploying, and operating large-scale parallel file systems. We believe that these lessons will be useful to the wider HPC community.
Storage backends of parallel compute clusters are still based mostly on magnetic disks, while newer and faster storage technologies such as flash-based SSDs or non-volatile random access memory (NVRAM) are deployed within compute nodes. Including these new storage technologies into scientific workflows is unfortunately today a mostly manual task, and most scientists therefore do not take advantage of the faster storage media. One approach to systematically include nodelocal SSDs or NVRAMs into scientific workflows is to deploy ad hoc file systems over a set of compute nodes, which serve as temporary storage systems for single applications or longer-running campaigns. This paper presents results from the Dagstuhl Seminar 17202 “Challenges and Opportunities of User-Level File Systems for HPC” and discusses application scenarios as well as design strategies for ad hoc file systems using node-local storage media. The discussion includes open research questions, such as how to couple ad hoc file systems with the batch scheduling environment and how to schedule stage-in and stage-out processes of data between the storage backend and the ad hoc file systems. Also presented are strategies to build ad hoc file systems by using reusable components for networking and how to improve storage device compatibility. Various interfaces and semantics are presented, for example those used by the three ad hoc file systems BeeOND, GekkoFS, and BurstFS. Their presentation covers a range from file systems running in production to cutting-edge research focusing on reaching the performance limits of the underlying devices.
Parallel input/output characterization studies and experiments with flexible resource management algorithms indicate that adaptivity is crucial to file system performance. In this paper we propose an automatic technique for selecting and refining file system policies based on application access patterns and execution environment. An automatic classification framework allows the file system to select appropriate caching and pre-fetching policies, while performance sensors provide feedback used to tune policy parameters for specific system environments. To illustrate the potential performance improvements possible using adaptive file system policies, we present results from experiments involving classification-based and performance-based steering.
Techniques are provided for globally appending data from a group of distributed processes to a shared file using a log-structured file system. Data generated by a plurality of processes in a parallel computing system are appended to a shared file by storing the data to the shared file using a log-structured file system (such as a Parallel Log-Structured File System (PLFS)); and generating an index entry for the data, the index entry comprising a logical offset entry and a timestamp entry indicating a time of the storage, wherein the logical offset entry is resolved at read time. The logical offset entry can be populated with an append placeholder that is resolved when the shared file is read. At read time, a plurality of the index entries associated with the shared file can be sorted using the timestamp entry to deliver the requested shared file to a requesting application.
When deployed in 2008/2009 the Spider system at the Oak Ridge National Laboratory s Leadership Computing Facility (OLCF) was the world s largest scale Lustre parallel file system. Envisioned as a shared parallel file system capable of delivering both the bandwidth and capacity requirements of the OLCF s diverse computational environment, Spider has since become a blueprint for shared Lustre environments deployed worldwide. Designed to support the parallel I/O requirements of the Jaguar XT5 system and other smallerscale platforms at the OLCF, the upgrade to the Titan XK6 heterogeneous system will begin to push the limits of Spider s original design by mid 2013. With a doubling in total system memory and a 10x increase in FLOPS, Titan will require both higher bandwidth and larger total capacity. Our goal is to provide a 4x increase in total I/O bandwidth from over 240GB=sec today to 1TB=sec and a doubling in total capacity. While aggregate bandwidth and total capacity remain important capabilities, an equally important goal in our efforts is dramatically increasing metadata performance, currently the Achilles heel of parallel file systems at leadership. We present in this paper an analysis of our current I/O workloads, our operational experiences with the Spider parallel file systems, the high-level design of our Spider upgrade, and our efforts in developing benchmarks that synthesize our performance requirements based on our workload characterization studies.
This paper shows mathematically and experimentally that inline deduplication is not suitable for file systems on ultra-low latency Intel Optane DC PM devices in terms of performance, and proposes DeNova, an offline deduplication specially designed for log-structured NVM file systems such as NOVA. DeNova offers high-performance and low-latency I/O processing and executes deduplication in the background without interfering with foreground I/Os. DeNova employs DRAM-free persistent deduplication metadata, favoring CPU cache line, and ensures failure consistency on any system failure. We implement DeNova in the NOVA file system. Evaluation with DeNova confirms a negligible performance drop of baseline NOVA of less than 1%, while gaining high storage space savings. Extensive experiments show DeNova is failure consistent in all failure scenario cases.
Many scientific applications have intense computational and I/O requirements. Although multiprocessors have permitted astounding increases in computational performance, the formidable I/O needs of these applications cannot be met by current multiprocessors a their I/O subsystems. To prevent I/O subsystems from forever bottlenecking multiprocessors and limiting the range of feasible applications, new I/O subsystems must be designed. The successful design of computer systems (both hardware and software) depends on a thorough understanding of their intended use. A system designer optimizes the policies and mechanisms for the cases expected to most common in the user's workload. In the case of multiprocessor file systems, however, designers have been forced to build file systems based only on speculation about how they would be used, extrapolating from file-system characterizations of general-purpose workloads on uniprocessor and distributed systems or scientific workloads on vector supercomputers (see sidebar on related work). To help these system designers, in June 1993 we began the Charisma Project, so named because the project sought to characterize 1/0 in scientific multiprocessor applications from a variety of production parallel computing platforms and sites. The Charisma project is unique in recording individual read and write requests-in live, multiprogramming, parallel workloads (rather than from selected or nonparallel applications). In this article, we present the first results from the project: a characterization of the file-system workload an iPSC/860 multiprocessor running production, parallel scientific applications at NASA's Ames Research Center.
Distributed burst buffers are a promising storage architecture for handling I/O workloads for exascale computing. As they are being deployed on more supercomputers, a file system that efficiently manages these burst buffers for fast I/O operations carries great consequence. Over the past year, FSU team has undertaken several efforts to design, prototype and evaluate distributed file systems for burst buffers on HPC systems. These include MetaKV: a Key-Value Store for Metadata Management of Distributed Burst Buffers, a user-level file system with multiple backends, and a specialized file system for large datasets of deep neural networks. Our progress for these respective efforts are elaborated further in this report.
Scientific and big data computations are increasingly being distributed across wide-area networks, and they often require access to remote files. The file systems that are directly mounted over wide-area networks transparently support such computations, and also obviate the need for special purpose file transfer tools. In typical distributed file systems, the access is limited to local sites, and in particular, the reach of Lustre file system implemented over InfiniBand (IB) is limited to at most tens of miles due to 2.5ms latency bound. We describe LNet router methods that connect IB Lustre file system to remote Ethernet clients over wide-area networks. We collect extensive Lustre throughput measurements over 10Gbps connections with 0-366ms round-trip times. They demonstrate that Gbps throughput can be sustained over connections spanning the globe. We present Lustre throughput profiles over local and wide-area connections, which show the effects of various buffers and credits; in particular, they highlight the throughput limits for large transfers over wide-area connections. Furthermore, the measurements show the positive effects of pipelining in achieving higher throughput for successively file transfers compared to rates indicated by IOzone benchmark rates.
AFS(originally, Andrew File System) is a widely-deployed distributed file system product used by companies, universities, and laboratories world-wide. However, it is not trivial to operate: runing an AFS cell is a formidable task. It requires a team of dedicated and experienced system administratores who must manage a user base numbring in the thousands, rather than the smaller range of 10 to 500 faced by the typical system administrator.
The approach has been to develop Parallel Virtual File System version 2 (PVFS2) , retaining the basic philosophy of the original file system but completely rewriting the code. It shows the architecture of the server and client components. BMI - BMI is the network abstraction layer. It is designed with a common driver and modules for each protocol supported. The interface is non-blocking, and provides mechanisms for optimizations including pinning user buffers. Currently TCP/IP and GM(Myrinet) modules have been implemented. Trove -Trove is the storage abstraction layer. It provides for storing both data spaces and name/value pairs. Trove can also be implemented using different underlying storage mechanisms including native files, raw disk partitions, SQL and other databases. The current implementation uses native files for data spaces and Berkeley db for name/value pairs.
Improved techniques are provided for storing metadata associated with a plurality of sub-files associated with a single shared file in a parallel file system. The shared file is generated by a plurality of applications executing on a plurality of compute nodes. A compute node implements a Parallel Log Structured File System (PLFS) library to store at least one portion of the shared file generated by an application executing on the compute node and metadata for the at least one portion of the shared file on one or more object storage servers. The compute node is also configured to implement a partitioned data store for storing a partition of the metadata for the shared file, wherein the partitioned data store communicates with partitioned data stores on other compute nodes using a message passing interface. The partitioned data store can be implemented, for example, using Multidimensional Data Hashing Indexing Middleware (MDHIM).
NERSC's newest system, Perlmutter, features a 35 PB all-flash Lustre file system built on HPE Cray ClusterStor E1000. Here, we present its architecture, early performance figures, and performance considerations unique to this architecture. We demonstrate the performance of E1000 OSSes through low-level Lustre tests that achieve over 90% of the theoretical bandwidth of the SSDs at the OST and LNet levels. We also show end-to-end performance for both traditional dimensions of I/O performance (peak bulk-synchronous bandwidth) and nonoptimal workloads endemic to production computing (small, incoherent I/Os at random offsets) and compare them to NERSC's previous system, Cori, to illustrate that Perlmutter achieves the performance of a burst buffer and the resilience of a scratch file system. Finally, we discuss performance considerations unique to all-flash Lustre and present ways in which users and HPC facilities can adjust their I/O patterns and operations to make optimal use of such architectures.
The new and substantially upgraded version of the Table File System in MIDAS is presented as a scientific database system. MIDAS applications for performing database operations on tables are discussed, for instance, the exchange of the data to and from the TFS, the selection of objects, the uncertainty joins across tables, and the graphical representation of data. This upgraded version of the TFS is a full implementation of the binary table extension of the FITS format; in addition, it also supports arrays of strings. Different storage strategies for optimal access of very large data sets are implemented and are addressed in detail. As a simple relational database, the TFS may be used for the management of personal data files. This opens the way to intelligent pipeline processing of large amounts of data. One of the key features of the Table File System is to provide also an extensive set of tools for the analysis of the final results of a reduction process. Column operations using standard and special mathematical functions as well as statistical distributions can be carried out; commands for linear regression and model fitting using nonlinear least square methods and user-defined functions are available. Finally, statistical tests of hypothesis and multivariate methods can also operate on tables.
This is a case study. It deals with the use of a 'virtual file system' (VFS) for Boeing's UNIX-based Product Standards Data System (PSDS). One of the objectives of PSDS is to store digital standards documents. The file-storage requirements are that the files must be rapidly accessible, stored for long periods of time - as though they were paper, protected from disaster, and accumulative to about 80 billion characters (80 gigabytes). This volume of data will be approached in the first two years of the project's operation. The approach chosen is to install a hierarchical file migration system using optical disk cartridges. Files are migrated from high-performance media to lower performance optical media based on a least-frequency-used algorithm. The optical media are less expensive per character stored and are removable. Vital statistics about the removable optical disk cartridges are maintained in a database. The assembly of hardware and software acts as a single virtual file system transparent to the PSDS user. The files are copied to 'backup-and-recover' media whose vital statistics are also stored in the database. Seventeen months into operation, PSDS is storing 49 gigabytes. A number of operational and performance problems were overcome. Costs are under control. New and/or alternative uses for the VFS are being considered.
This document discusses the performance and advantages of striped file systems on the SGI AD workstations. Performance of several striped file system configurations are compared and guidelines for optimal striping are recommended.