Agencies help test cloud-based file storage system
Appistry system could help speed the handling of large volumes of data
- By Rutrell Yasin
- Mar 16, 2010
Federal agencies and systems integrators are beta testing a cloud-based file storage system from Appistry that distributes files and requests across multiple machines, allowing servers to be used for both computational purposes and for file storage. The system could help avoid bottlenecks and performance degradation during the analysis, processing and storage of large volumes of data-intensive applications.
Typically, those computing resources are in different locations in data centers, said Kevin Jackson, engineering fellow at NJVC, a federal systems integrator that is testing Appistry’s new CloudIQ Storage system.
NJVC provides IT solutions to defense and intelligence agencies, most significantly to the National Geospatial-Intelligence Agency.
Don't look down: The path to cloud computing is still missing a few steps
Cloud computing has appeal for Web applications
CloudIQ Storage is “the beginning of a change in the approach to computational storage, which takes advantage of what you can do in the cloud,” Jackson said.
Computational storage is a concept that unifies applications and data by storing data across commodity servers and then migrating application processing to machines that contain the relevant data.
“With cloud computing you can do things in parallel as long as you are smart about where you put data and where you store data,” Jackson said.
Traditional application architectures rely on the transfer of data files over expensive and bandwidth-constrained networks processed by application logic, Appistry officials said.
CloudIQ Storage is architected for petabyte-scale data, providing a fully distributed architecture with no single point of failure or network bottlenecks, they said.
CloudIQ Storage can be used as a stand-alone cloud storage system or in conjunction with Appistry CloudIQ Engine for computational storage. The storage system is an extension of the company’s Cloud IQ enterprise cloud platform.
NJVC last year used that platform to cloud-enable a client /server application for a federal customer in about three weeks, Jackson noted.
CloudIQ Storage is being tested in a NJVC data center “out west and we are cautiously optimistic,” about the system’s potential, Jackson said.
CloudIQ Storage is expected to be available this spring.
Appistry also announced a version of CloudIQ Storage that supports Apache Hadoop, an open-source project for processing and analyzing large amounts of data. The Hadoop project provides MapReduce, a software framework for distributed processing of large datasets, and Hadoop Distributed File System designed to support MapReduce.
Hadoop and MapReduce are becoming popular among organizations that need to analyze large quantities of data, Appistry officials said.
Rutrell Yasin is is a freelance technology writer for GCN.