ETD

• auto-scaling
• big data
• cep
• cloud
• Complex Event Processing
• elasticity

One key topic in cloud computing is elasticity, which is the ability of the cloud environment to timely adapt the resource assignment along with the workload demand. According
to cloud on-demand model, the infrastructure should be able to scale up and down to unpredictable workloads, in order to achieve both a guaranteed service level and cost efficiency.

This work addresses the cloud elasticity problem, with particular reference to the Complex
Event Processing (CEP) systems.
CEP systems are designed to process large volumes of event-driven data streams and
continuously provide results with a low latency and in real-time. CEP systems need to
adapt to changing query and events loads. Because of the high computational requirements
and varying loads, CEP are distributed system and running on cloud infrastructures.

In this work we review the cloud computing auto-scaling solutions, and study their suit-
ability in the CEP model. We implement some solutions in a CEP prototype and evaluate
the experimental results.