Database Reference
In-Depth Information
Time Series Data
In this chapter, we will cover the following topics:
• What we mean by time series data
• The special design challenges that time series data presents
• Which important performance considerations apply
• The best practices to consider when dealing with time series data
All of us are familiar with time series data. Although we will provide a definition
and examples later, you can think of it as minute-by-minute trading data. Such data
presents a number of challenges for the HBase designer, as they need to create a
proper schema for the data, find a balance between convenience and performance,
and also keep factors such as as the overheating of specific region servers and
bloated bloom filters in mind.
If all of this sounds obscure at the moment, we will try to make it clear by the
end of this chapter. As we will see, HBase is perfect for storing time series data.
However, one needs to pay attention to a number of pesky details.
Let's start with a definition of time series data. Time series data is a type of data that
is recorded at regular time intervals. Some examples of time series data include logs,
sensor data (power grid), stock ticks, monitoring systems, and many others.
Please note that
logs
can have a few different meanings. In the modern world, when
talking about logs, we are usually referring to web logs or the recording of events
that happens when users browse a website and a web server sends them the pages
and records their actions.
