Distributed Cache and CEP with Storm - Real-time Analytics with Storm and Cassandra - page 240

Esper
Esper is one of the leading CEP engines that are available under open source—GPL and
Enterprise License. The package can be downloaded from http://www.espertech.com/
download/ , and if you are attempting to execute a Maven-based Esper project, the depend-
ency can be built as follows:
<dependency>
<groupId>com.espertech</groupId>
<artifactId>esper</artifactId>
<version> ... </version>
</dependency>
Ref :Espertech.com
The next obvious question could be why we want to use Esper-CEP in conjunction with
Storm. Well, Esper has some unique capabilities that work well with Storm and let the EQL
facility leverage the results drawn over Storm. The following are complementing features
that lead to this choice:
• Throughput : Complementing the capability of Storm, Esper also has a very high
throughput and can process from 1K to 100K messages per second.
• Latency : Esper has the ability to perform EQLs and actions based on results of
Esper at a very low latency rate; in most cases, this is the order of milliseconds.
• Computations : These refer to the ability to perform functions such as pattern de-
tection, complex queries based on aggregates, and correlation over time. These
slice windows of streaming data.