Conference Paper

A Big Picture for Big Data

Abstract: 

The omnipresent data deluge commonly is characterized by the four "Big Data" Vs, Volume, Velocity, Variety, and Veracity. Commonly, most emphasis today is put on Volume, and to a lesser extent on Velocity.
Many disciplines, however, like oil and gas industry for example, emphasize that the Variety is their main challenge: a plethora of data structures, data formats, ways of using data. One attempt to capture this is a common integrating architecture, like the NIST Big Data Reference Architecture; this is contrasted by statements like "no one size fits all".

In this talk we attempt to classify core data structures that commonly appear in geo science and engineering, and elsewhere. Such common data structures can be supported by specific models and systems. If we follow this approach then the next quest obviously is how to (re) integrate them.
We argue that a new generation of mediating technology, allowing both logical and physical distribution in a new sense, is required. The case we make is backed by observations from various earth sciences, standardization directions, and information system architectures.

Author: 

Peter Baumann

Presenter Biography: 

Peter Baumann is Professor of Computer Science at Jacobs University Bremen where he is researching on large-scale multi-dimensional array databases, generally on "Big Data in Science".
Peter Baumann has coined the research field of Array Databases by introducing rigorous methods for storage and query support of massive multi-dimensional arrays in databases. This includes establishing the formal foundations through Array Algebra, architecting the worldwide first complete Array DBMS, rasdaman, applying rasdaman in Earth, Space, and Life Sciences, and - based on this experience - standardizing access interfaces for array query languages in OGC, ISA, and RDA.

? Top