Pluggable Event Architecture for Cloud Environments

Project Information

Discipline
Computer Science (401) 
Orientation
Research 
Abstract

In this experiment we seek to better understand the use of Complex Event Processing (CEP) to scale systems deployed in cloud environments. The work is born out of a CEP model used as part of the One Degree Imager Portal, Pipeline and Archive (ODI-PPA) (http://uits.iu.edu/page/bblr). Specifically the CEP model has been deployed within the ODI-PPA data subsystem which is responsible for data archive/retrieval and workflow execution. The model has demonstrated a significant reduction in system coupling and appears to afford a high degree of horizontal scalability. We expect that this experiment will provide evidence of the models high degree of scalability within cloud environments.

Intellectual Merit

This experiment hopes to demonstrate a unique model for scaling multi-component systems within a cloud environment. Furthermore, in the pursuit of demonstrating the model at scale this experiment will also demonstrate the model's use in the ingest, archive and retrieval of large data sets. Thus establishing merit within the criteria of upgrading the computation and computing infrastructure.

Broader Impacts

It is expected that the broader impact of this work will be in the NSF merit review criteria of : Upgrade the computation and computing infrastructure, including advanced computing resources and new types of information tools (e.g., large databases, networks and associated systems, and digital libraries).

Project Contact

Project Lead
Jeffrey Cox (jefcox) 
Project Manager
Jeffrey Cox (jefcox) 

Resource Requirements

Hardware System
  • Not sure
 
Use of FutureGrid

The intended use of FutureGrid is the ability to provision VM's within a cloud environment and test the given model at scale.

Scale of Use

It will be necessary to: 1) test and debug the model within the experimental system. This will require only a few VM's at a time. 2) once the experimental system is stable, we will test the system as we increase the number of VM/components involved. This will require a much larger number of VM's than in 1, but only for the duration of the tests. It is expected that the tests will take only a few hours at each scale measurement.

Project Timeline

Submitted
12/11/2012 - 17:05 
»