High Performance Clustered Storage for Image Management

描述(由申请人提供)：项目摘要/摘要我们建议购买并运营一个高性能存储集群，作为波士顿地区多个机构的研究人员的共享资源。该存储集群将根据越来越多从事高通量实时和固定细胞成像、基于细胞的筛选(“高内容筛选”)和自动图像分析(“机器视觉”)的显微镜工作者的需求进行定制。具体项目包括创建以电子形式捕捉斑马鱼发育历史的“数字鱼”，对正常细胞和转化细胞中的信号转导电路进行快速细胞分析，对微管蛋白和肌动蛋白细胞骨架进行散斑显微镜检查，以及大规模RNAi筛选参与苍蝇发育的新基因。提议的存储群集的主要用户目前依赖于HMS IT提供的通用网络连接存储，该存储缺乏下一代映像项目所需的容量和性能。由EMC Celerra NAS组成的现有基础架构已达到其绝对设计极限，扩展起来过于繁琐且成本高昂。相比之下，建议的120 TB IQ12000x-EX存储群集(来自西雅图西雅图Isilon Inc.)代表了复杂性和成本之间的理想折衷，并且可以扩展到&gt；3 PB。Isilon通过将标准硬件链接到高度优化的专有操作系统和软件堆栈来实现其低成本和高性能，这些操作系统和软件堆栈执行文件系统、卷管理器和RAID管理器的功能。Isilon群集中的高级功能将使多个位置的用户能够在直观、无缝的用户环境中存储和管理图像数据。该产品的非凡简单性还将减轻HMS IT人员、科学家和存储经理的现有负担。推荐的Isilon群集的总体好处是，它将满足主要用户对更多基于磁盘的计算机存储的迫切需求，显著提高这些用户管理可用空间的能力，并支持不断增长的新用户群体。数据存储和处理对于光学显微镜和图像分析的创新越来越重要。因此，Isilon群集的科学基础是强有力的，而且需求迫在眉睫。此外，这台机器是由一个目前正在经历快速需求下降的行业在美国设计和制造的；因此，我们将帮助维持一个对科学计算未来至关重要的经济部门的美国就业。 公共卫生相关性：大规模活细胞显微镜和固定细胞显微镜在细胞生物学、基因鉴定和药物发现中发挥着越来越重要的作用。下一代成像项目依赖于管理和处理数字数据的计算机系统，就像依赖用于收集数据的显微镜一样。本提案中要求的复杂、可扩展和美国制造的集群式计算机存储系统对于在全国以其在显微镜方面的专业知识而闻名的用户群体的实验室的持续创新以及满足大量新接触数字显微镜的研究人员日益增长的需求是必要的。

DESCRIPTION (provided by applicant): Project Summary/Abstract We propose to purchase and operate a high performance storage cluster as a shared resource for investigators at multiple institutions in the Boston area. The storage cluster will be customized to the needs of a growing community of microscopists involved in high throughput live and fixed-cell imaging, cell-based screening ("high content screening") and automated image analysis ("machine vision"). Specific projects include creation of a "digital fish" that captures in electronic form the development history of Zebrafish, rapid cell-based analysis of signal transduction circuits in normal and transformed cells, speckle microscopy of the tubulin and actin cytoskeleton and large-scale RNAi screening for new genes involved in fly development. The major users of the proposed storage cluster currently rely on a general purpose, network attached storage provided by HMS IT that lacks the capacity and performance needed for next-generation imaging projects. The existing infrastructure, consisting of EMC Celerra NAS has reached its absolute design limits and is too cumbersome and expensive to expand. In contrast, the proposed 120 TB IQ12000x-EX Storage Cluster (from Isilon Inc, Seattle WA) represents the ideal tradeoff between sophistication and cost, and is expandable to >3 petabytes. The Isilon achieves its low cost and high performance by linking standard hardware to proprietary, highly optimized operating systems and software stacks that perform the funcitons of a filesystem, volume manager, and RAID manager. The advanced features in the Isilon cluster will make it possible for users at multiple locations to store and manage image data within an intuitive and seamless user environment. The remarkable simplicity of the product will also ease the existing burden on staff scientists and storage managers in HMS IT. The overall benefit of the proposed Isilon cluster is that it will address an acute requirement for more disc-based computer storage among major users, increase dramatically the ability of these users to manage the space that is available and support a growing community of new users. Data storage and processing are increasingly critical for innovation in optical microscopy and image analysis. The scientific rationale for the Isilon Cluster is therefore strong and the need immediate. Moreover, the machine is designed and manufactured in the United States by an industry that is currently experiencing rapidly falling demand; thus, we will help to sustain a US- based employment in a sector of the economy that is critical to the future of scientific computing. PUBLIC HEALTH RELEVANCE: Large scale live and fixed cell microscopy is playing an increasingly important role in cell biology, gene identification and drug discovery. Next generation imaging projects are as dependent on the computer systems that manage and process digital data as they are on the microscopes used to collect data. The sophisticated, expandable and US-manufactured clustered computer storage system requested in this proposal is necessary for continued innovation in the laboratories of a user group known nationally for its expertise in microscopy, and to meet the growing needs of a large community of investigators new to digital microscopy.