This approach is much faster than the traditional OLTP design of; Insert the row(s) into the database while constantly polling the data to see if something relevant has happened.
It is especially suited to situations where you have high speed inputs, a fixed set of fairly simple queries and may not need to keep all the data.
[citation needed] e.g.: Some industries measure the voltage, current and other attributes of hundreds of electric motors in their conveyor belts, 100 times each second.
Then compare each measurement to its average, plant operators are alerted should a sudden change occur.
StreamInsight is a separate install and not visible as a part of a regular SQL Server setup.