|
Activity Number:
|
498
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 5, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Risk Analysis
|
| Abstract - #302799 |
|
Title:
|
Statistical Inference and Simulation on Security Metrics in Cloud Computing for Large Cyber Systems
|
|
Author(s):
|
Luis A. Cueva-Parra*+ and Mehmet Sahinoglu and David Tyson and Sunil Das
|
|
Companies:
|
Auburn University Montgomery and Auburn University Montgomery and Auburn University Montgomery and Troy University Montgomery
|
|
Address:
|
P.O. Box 244023, Montgomery, AL, 36124-3127
36124,
|
|
Keywords:
|
CLOUD ; compound Poisson process ; loss of service ; security ; simulation
|
|
Abstract:
|
This paper studies the service quality of a conglomerate of distributed parallel computers (an example of a CLOUD) working independently or with a definable dependence relationship, and additively to serve the needs of a large number of customers in demand. Statistical Inference on Loss of Service (LoS) in the event of security breaches as a metric is conducted to further see what risk management must be planned to improve the service quality. We will propose a mathematical-statistical theoretical model to describe the indices for availability (presence of security) and unavailability (lack of security) of the entire system operation assuming that the failure and repair count of a node is governed by a Compound Poisson process. Both theoretical analysis and discrete-event simulation models using JAVA code will be complemented by a set of experiments in simulated and real systems.
|
