Complex Event Processing (CEP) is a powerful paradigm that can derive correlations from different data sources for a wide variety of applications. CEP provides semantic units called operators e.g., filter and join, that collectively represent a complex event. In current CEP systems, operators are highly dependent on the programming language and the underlying server. This restricts the capability of provisioning user-defined operators at runtime as well as the flexibility of developing server agnostic custom operators. In this paper, we provide a serverless CEP architecture, which offers developers the flexibility to design operators in any language and integrate them at runtime. We embed operators in the function as a service model of serverless architecture. This is very beneficial for applications such as financial fraud detection where complex machine learning operators must be integrated at runtime to avoid service disruption. We show using our preliminary evaluation that only with minimal overhead in latency, we can offer highly flexible server agnostic CEP operators.
[1]
Schahram Dustdar,et al.
A Serverless Real-Time Data Analytics Platform for Edge Computing
,
2017,
IEEE Internet Computing.
[2]
Alessandro Margara,et al.
Processing flows of information: From data stream to complex event processing
,
2012,
CSUR.
[3]
Jignesh M. Patel,et al.
Twitter Heron: Stream Processing at Scale
,
2015,
SIGMOD Conference.
[4]
Seif Haridi,et al.
Apache Flink™: Stream and Batch Processing in a Single Engine
,
2015,
IEEE Data Eng. Bull..
[5]
Boris Koldehofe,et al.
TCEP: Adapting to Dynamic User Environments by Enabling Transitions between Operator Placement Mechanisms
,
2018,
DEBS.