Database Reference
In-Depth Information
Cloud Analytics
Google's cloud analytics services enable you to extract meaning from your
data:
•
Cloud SQL
: A hosted MySQL instance in the cloud
•
Prediction API
: Enables you to train machine learning models and
apply them to your data
•
Cloud Hadoop
: Packages Hadoop and makes it easy to run on Google
Compute Engine
•
BigQuery
: Enables you to run SQL statements over your structured
data
If you find that something is missing from Google's Cloud Platform, you
always have the option of running your favorite open source software stack
on Google Compute Engine. For example, the Google Cloud Hadoop
package is one way of running Hadoop, but if you want to run a different
version of Hadoop than is supported, you can always run Hadoop directly;
Google's Hadoop package uses only publicly available interfaces.
Problem Statement
Before we go on to talk about BigQuery, here's a bit of background
information about the problems that BigQuery was developed to solve.
What Is Big Data?
There are a lot of different definitions from experts about what it means
to have Big Data; many of these definitions conceal a boast like, “Only a
petabyte? I've forgotten how to count that low!” This topic uses the term Big
Data to mean more data than you can process sequentially in the amount
of time you're willing to spend waiting for it. Put another way, Big Data
just means more data than you can easily handle using traditional tools
such as relational databases without spending a lot of money on specialized
hardware.
This definition is deliberately fuzzy; to put some numbers behind it, we'll
say a hundred million rows of structured data or a hundred gigabytes of
unstructured data. You can fit data of that size on a commodity disk and
even use MySQL on it. However, dealing with data that size isn't going to
