Evaluating the performance of the dNFSP file system

Abstract

Parallel I/O in cluster computing is one of the most important issues to be tackled as clusters grow larger and larger. Many solutions have been proposed for the problem and, while effective in terms of performance, they usually represent a considerable amount of hacking into a "traditional" Beowulf cluster installation. In this paper, we investigate a parallel solution based on NFS, which reduces the level of intrusion in the file server installation, keeps the client side untouched, and still provides an improved level of performance and scalability for parallel applications. We compare our proposal to other existing file systems using known benchmarks, and demonstrate that it is a valid alternative for general-purpose cluster computing.

Impact of I/O Coordination on a NFS-Based Parallel File System with Dynamic Reconfiguration

“…Clients mount a volume in one metaserver, so each meta-server handles only a subset of the clients. In dNFSp, if all the clients access the file system at the same time, the load will be distributed among these servers [12].…”

Amphidinolide B: Asymmetric Synthesis of a C7−C20 Synthon

“…Aldehyde 8 participated in asymmetric acyl halide-aldehyde cyclocondensation (AAC) with acetyl bromide using the Al(III) catalyst 9 (10 mol %) to afford the 1,3- syn β-lactone 10 and correctly established the C 18 stereocenter (88%, 94% de) (Scheme ). , Amine-mediated lactone ring opening and protection of the emergent hydroxyl function as the tert -butyldimethylsilyl ether provided Weinreb amide 11 (86% two steps). Selective cleavage of the alkynyl silane was achieved using KOH/MeOH to afford terminal alkyne 12 (93%), thereby providing the conduit for stereoselective installation of the C 14 −C 15 trisubstituted olefin.…”

Metaserver Locality and Scalability in a Distributed NFS

“…In previous works [2,3], we have presented the dNFSp file system, an extension of NFSv2 that aims at improving both performance and scalability of a regular NFS server while keeping its standard administration procedures and, mainly, compatibility with the regular NFS clients available on every Unix system. Similarly to other parallel file systems such as PVFS [4] and Lustre [5], dNFSp is based on a distributed approach where the gain in performance is obtained by executing tasks in parallel over several machines of the cluster.…”