Matrix Computations
- PDF / 3,520,823 Bytes
- 159 Pages / 565.087 x 755.008 pts Page_size
- 45 Downloads / 260 Views
MapReduce Massive-Scale Analytics
MasPar Synonyms SIMD (Single Instruction, Multiple Data) Machines MASPAR [] were SIMD machines built by MasPar Computer Corporation. The first system was delivered in .
Related Entries Flynn’s Taxonomy
Bibliography . Tom Blank () Compcon Spring ’: the MasPar MP- Architecture, San Francisco, CA, USA, Feb– Mar , pp –
Massively Parallel Processor (MPP) Distributed-Memory Multiprocessor
Massive-Scale Analytics Amol Ghoting , John A. Gunnels , Mark S. Squillante IBM Thomas. J. Watson Research Center, Yorktown Heights, NY, USA IBM Corp., Yorktown Heights, NY, USA IBM, Yorktown Heights, NY, USA
Synonyms Deep analytics; Large-scale analytics; MapReduce
Definition Massive-scale analytics refers to a combination of mathematical methods and high-performance computational methods for the analysis of vast amounts of data in order to gain crucial insights and facilitate decision making across a broad spectrum of application domains, sometimes in an automated data-driven fashion. This requires a unified approach that addresses the two key dimensions of massive-scale analytics: data and computation.
Discussion Introduction The volumes of data available to various organizations throughout society over the past many years have been growing at an explosive rate. Important advances in computer sciences and technologies have made this possible, which in turn have resulted in the development and growth of data warehouses and reporting capabilities to view information concerning various aspects of an enterprise. These basic reporting capabilities have helped organizations improve the accuracy and timeliness of the data used for the purposes of enterprise decision making. However, significantly increasing the rewards for investments in data collection and data warehouse construction well beyond basic reporting requires the development of massive-scale analytics that provide deeper insights and understanding of the enterprise and optimization of its performance. The overall goal of massive-scale analytics is to enable organizations with data-driven decision-making capabilities based on a combination of advanced mathematical and computational methods, including automated tools and applications. There are many challenges to realizing the full benefits of massive-scale analytics. This includes important challenges along the two key dimensions of applying massive-scale analytics against huge volumes of data, namely the data dimension which involves machine
David Padua (ed.), Encyclopedia of Parallel Computing, DOI ./----, © Springer Science+Business Media, LLC
M
Massive-Scale Analytics
learning, knowledge discovery, and data management, together with the computational dimension which involves high-performance and massively parallel computing. The main challenges for the data dimension entail extracting knowledge and insight from complex queries against massive data sets in an efficient and effective manner. The main challenges for t
Data Loading...
