Resumen
In spite of all the advantages delivered by cloud computing, several challenges are hindering the migration of customer software and data into the cloud. On top of the list is the security and privacy concerns arising from the storage and processing of sensitive data on remote machines that are not owned, or even managed by the customers themselves. In this paper, initially a homomorphic encryption-based Cryptographic Agent is proposed. The proposed Cryptographic Agent is based on Paillier scheme, and is supported by user-configurable software protection and data privacy categorization agents, as well as set of accountable auditing services required to achieve legal compliance and certification. This scheme was tested using different text documents with different sizes. Testing results showed that as the size of the document increases, the size of the generated key increases dramatically causing a major problem in regards to the processing time and the file size especially for large documents. This leaded us to the second part of this research which is: a modified security architecture that adds two major autonomic security detective agents to the multi-agent architecture of cloud data storage. In this paper, we focus on the first agent namely (Automated Master Agent, AMA) that is added to the Multi Agent System Architecture (MASA) layer (cloud client-side) by which any changes happen in the document are mapped in a QR code encoded key print (KP). Experimental results after integrating these agents showed a 100% alternation detection accuracy and a superiority in extracting the KP of large and very large size documents which exceeds the currently available products and leverage the tamper-proof capabilities of cryptographic coprocessors to establish a secure execution domain in the computing cloud that is physically and logically protected from unauthorized access.