Think Global, Act Local: A Buffer Cache Design for Global Ordering and Parallel Processing in the WAFL File System

2016 ◽  
Author(s):  
Peter R. Denz ◽  
Matthew Curtis-Maury ◽  
Vinay Devadas
Keyword(s):  
Parallel Processing ◽  
File System ◽  
Buffer Cache ◽  
Cache Design

scholarly journals Considering Data Reference Pattern in Buffer Cache for Continuous Media File System

2002 ◽  
Vol 9A (2) ◽  
pp. 163-170
Author(s):  
Kyung-Woon Cho ◽  
Yeon-Seung Ryu ◽  
Kern Koh
Keyword(s):  
File System ◽  
Continuous Media ◽  
Reference Pattern ◽  
Buffer Cache

scholarly journals Rasterhadoop: An Application Perspective of Raster Data Processing on Hadoop

2019 ◽  
Vol 8 (4) ◽  
pp. 11147-11150
Keyword(s):  
Remote Sensing ◽  
Image Processing ◽  
Parallel Processing ◽  
Data Processing ◽  
Satellite Data ◽  
Spatial Data ◽  
File System ◽  
Distributed File System ◽  
Raster Data ◽  
Volume Velocity

Hadoop is currently the most popular platform for parallel processing. With its two major components namely the Distributed File System (HDFS) and a parallel processing paradigm (MapReduce) in addition to its ease of installation and usage, Hadoop has become the chosen platform for efficiency whether in the commercial arena or the scientific arena such as Satellite Data Processing. The number of remote sensing satellites have also grown leaps and bounds and the data sent back by them for processing has all the three characteristics namely volume, velocity and variety that make it Big Spatial Data. In this paper, we present the extensions provided to Hadoop that enable Image Processing using legacy code and further elaborate on the various methods provided.

Educational Resources Cloud Platform Based on Hadoop

2014 ◽  
Vol 912-914 ◽  
pp. 1249-1253
Author(s):  
Qing Hou ◽  
Lei Pan ◽  
Jia Xi Xu ◽  
Kai Zhou
Keyword(s):  
Parallel Processing ◽  
File System ◽  
Educational Resources ◽  
Distributed File System ◽  
Distributed Programming ◽  
Massive Data ◽  
Mass Storage ◽  
Cloud Platform ◽  
Process Data ◽  
Programming Framework

As the traditional educational resources platform exists some deficiencies in storage, parallel processing, and cost, we designed a cloud platform of educational resources based on Hadoop framework.The platform applied HDFS distributed file system to store massive data distributedly and applied MapReduce distributed programming framework to process data parallelly and scheduler resources, which solved the mass storage of resources and improved the efficiency of various resources retrieval.

Sign in / Sign up

Export Citation Format

Share Document