Agents based Security Model for Cloud Database Applications
Abstract
Cloud computing offers a number of benefits but security challenges has remained a big barrier
for its widespread applications. As cloud computing moves the application software and
databases to be hosted by third parties, the management of such data and services becomes
untrustworthy. The security challenges include unauthorized access, reduced control over data,
over reliance on service providers and encryption performance degradation. This study sought to
explore security techniques in current research works; identify cloud security challenges amongst
users and cloud providers in Nairobi, Kenya; and develop and test an agent-based cloud security
model. Findings from different research works revealed encryption having performance
degradation issues, no real time data monitoring tool for clients, and data residing in one storage
location and can easily be accessed upon intrusion. Sample population from financial sector and
cloud service providers were used for the study. A multi-agent cloud security model was
developed, using the Prometheus multi-agent methodology, to provide a combined and
intelligent security solution. Research findings revealed concerns over cloud security and this
model provides the needed solution encompassing data classification and separation, selective
encryption, access control, use of agents to search and update massive volumes of data, and realreal
time monitoring. Data classification and separation allows data to be kept at different
location thereby making data incomplete upon intrusion whereas selective encryption minimizes
the volume of data to be encrypted thereby reducing encryption and decryption performance
degradation.