Database Reference
In-Depth Information
5
IBM's Enterprise Hadoop:
InfoSphere BigInsights
Few technologies
have gotten more buzz over the past few years than
Hadoop and NoSQL. Couple that with Big Data, and you have enough
hype to write a whole library of trendy technical topics. There's a reason for
all the excitement around these technologies. Traditional data storage and
analytics tools have not been cutting it when dealing with Big Data. From a
volume
perspective, many tools start being impractical when thresholds in
the dozens of terabytes are exceeded. Sometimes “impractical” means that
the technology simply won't scale any further, or that a tipping point is
reached in terms of how much time it takes to transfer a data set over the
network for processing. And other times, impractical means that although
the technology can scale, the licensing, administrative, and hardware costs
to deal with increased volume become unpalatable. From a
variety
perspec-
tive, traditional analytic tools only work well with structured data, which
represents, at most, 20 percent of the data in the world today. Finally, there
is the issue of how fast the data is arriving at your organization's door-
step—Big Data velocity—which we detail in the next chapter.
Considering the pressing need for technologies that can overcome the
volume and variety challenges for data at rest, it's no wonder that business
magazines and online tech forums alike are buzzing about Hadoop and
NoSQL. And it's not all talk either. The IT departments in most Fortune 500
companies have done some level of experimentation with Hadoop and the like.
85
