Databases Reference
In-Depth Information
Broadly speaking, workload scan be classified into three primary groups:
•
Online Transaction Processing (OLTP):
Transaction processing
is a mixed read-write workload that can be become very
write-intensive. OLTP requires low latency response, accesses
small amounts of data at one time, and has predictable access
patterns with few complex joins between different sets of data.
Streaming data processing and complex event processing
type of requirements are at the other extreme end of the OLTP
spectrum.
•
Business intelligence (BI):
Originally, this was viewed as a
combination of batch and on-demand reporting, later expanded
to include ad hoc query, dashboards, and visualization tools.
BI workloads are read-intensive, with writes usually done during
off-hours or in ways that don't compete with queries. While quick
response times are desired, they are not typically in the sub-second
range that OLTP requires. Data access patterns tend to be
unpredictable; they often involve reading a lot of data at one time,
and can have many complex joins.
Newer concepts like data discovery and exploratory analytics
are two other types of workloads where it not only becomes
read-intensive but also highly iterative.
•
Analytics:
Analytic workloads involve more extensive calculation
over data than BI. They are both compute-intensive and read-
intensive. They generally access entire data sets or a combination
of different data sets at one time prior to doing computations.
Most analytic workloads are done in a batch mode, with the
output used downstream via BI or other applications.
Relational databases have been the platform of choice for the above-defined three
workloads over the past two decades. As workloads grew larger and more varied, the
databases kept adding new features to improve performance. Over the last decade, data
volumes and complexity of data types have pushed the workloads past the capabilities of
almost all of these RDBMS.
Workload Characteristics
The workload characteristics we are going to discuss below apply more to general
database management principles; however, it is essential to understand these
characteristics in light of big data and analytics requirements.
What do we mean by workload? Table
4-1
outlines the characteristics of workload in
relation to constraints on database.
