Open Access Open Access  Restricted Access Subscription Access

Predictive Modeling of Service Level Agreement Parameters for Cloud Services

Seema Sunil Chowhan, Shailaja Shirwaikar, Ajay Kumar

Abstract


Cloud computing has emerged as an important paradigm in Information and Communication Technology space by enabling cost effective, on demand provisioning of elastic computing resources. With limited or almost negligible upfront investment, lots of organizations are attracted towards cloud, for outsourcing their computational needs. Service Level Agreements (SLA) between Cloud providers and the Cloud users are used to assure Quality of Service (QoS) which is one of the big issue that resists organization from availing cloud resources. SLA management is thus an important activity for Cloud providers as SLA violations may lead to contractual penalties and in turn loss of revenue and customer base. Managing SLA involves constant monitoring and controlling various SLA parameters. Therefore, it is desirable for providers to control possible violations before they happen by predicting the values of SLA parameters using the values continuously measured over a time period. We present a predictive modeling approach for predicting SLA violations, for throughput as the SLA parameter. Available datasets containing measurements on web services are used for training the prediction model.

Full Text:

PDF


DOI: http://dx.doi.org/10.47164/ijngc.v7i2.308