Databases Reference
In-Depth Information
Chapter 6
Implementation of Big Data Analytics
This chapter provides glimpses of implementation plans and related
challenges. Big Data Analytics may disrupt a major BI program. Do we
use Big Data Analytics to radically transform this organization or evolve it for
balanced growth? How do we establish a road map and ind initial pilots? How
do we evolve data governance to include considerations for Big Data?
6.1 Revolutionary, Evolutionary, or Hybrid
A typical Big Data Analytics implementation delivers three signiicant advan-
cements in performance. First, it can reduce latency by an order of magnitude,
providing accessibility to data in minutes or seconds as opposed to hours or days.
Second, it increases the capacity to store data by an order of magnitude, moving
from terabytes to petabytes. Third, it offers a much lower cost of acquisition and
operation. Because the architecture is typically built on commodity hardware and
requires fewer administrators, the cost, too, is reduced by an order of magnitude.
However, these implementations require a commitment to Big Data Analytics
and a strong desire to migrate from the current platform. What if we have already
invested a large IT budget in conventional BI? How far do we go in the irst
phase? Do we replace the current Data Warehouse architecture or augment it with
Big Data Analytics tools? Both approaches have obvious pros and cons. In this
section, I describe the three alternatives and discuss what would tilt us in one
direction or another for a speciic implementation.
Before I review the alternatives, let us irst place the current environment in
the context of the architecture described in Chapter 5 and understand how similar
or dissimilar the architectures are.
In a typical “traditional” architecture, we have a set of components for
ingesting data, a set of components for storing the data, and a set of components
Search WWH ::
Custom Search