Warehousing for low latency analytics
Date:
By: Philip Howard
Classification: White Paper
Latency is defined as the time that elapses
between a stimulus and the response to it. In
the context of analytics this could be taken to
mean the time taken between initiating a query
and receiving an answer or, if you are querying
data on a continuous basis, the time taken
between a change in the data and that change
being reflected in the result set.
In this paper we will first discuss some use
cases. While by no means an exhaustive list
this will give the reader some idea of the
range of different applications in which low latency
is required, as well as the extent of that
latency. As we move forward we expect low
latency requirements to become more and
more prevalent.
We will conclude by discussing the sort of features
to look for in a data warehouse that is
required to support low latency analytics, and
follow that with a discussion of how the Vertica
Analytic Database meets those requirements.