Semantic Similarity Model for Risk Assessment in Forming Cloud Computing SLAs

Abstract

Cloud computing has enabled users to access various resources and applications as a service and in return pay the provider only for the time for which they are used. Service Level Agreements (SLA) are formed between the user and provider to ensure that the required services and applications are delivered as expected. With the increase of public cloud providers, challenges such as availability, reliability, security, privacy and transactional risk demand detailed assessment during the formation of SLAs. This paper focuses on one subcategory of transactional risk while forming SLAs: namely, performance risk. We argue that performance risk assessment should be done by the user before entering into an SLA with a service provider. We propose to measure performance risk according to the specific context and assessment criteria with the aid of a semantic similarity model for the SLA requirement being negotiated in a cloud computing environment. We show through simulations that the performance risk analysis is more accurate using semantic similarity matching compared with analysis without semantic similarity matching.