IT Services Metrics
January 1, 2009 - May 1, 2012
Regarding information technology (IT) artifacts as services delivered to the end user (XaaS) is a paradigm shift that is fast changing the way management looks at the role of IT within an organization. IT Service development and maintenance which was in the past regarded as an integral part of an organization, has been outsourced or off shored by most companies to external consultants who provide onsite or hosted services. This outsourcing model is now being replaced by a new delivery model with organizations obtaining services from multiple vendors distributed globally – the service provider is itself virtualized on the cloud. Measuring the Quality of such services provided via the cloud is an interesting open problem. Traditional quality measures such as SERVQUAL (Parasuraman et al) are defined on a typical B2C model, namely, a consumer using a (generally) human provided service. The service quality is defined as a difference between the perceived and expected quality. In virtualized service delivery models, not only are humans not always the providers, but the model is B-to (virtualized service providers)-C. In this project, we want to determine the overall performance metric of IT services. Our other key focus is on the fact that even though the user interacts with a single service at the front, virtualization often means that the service is composed of many others, and each atomic service can depend on underlying backend applications, which in turn depend on resources such as databases, network, servers, storage etc. for their successful delivery. We will explore how to translate the hard SLA metrics of service components, applications and resources to overall performance of the Service delivered. Not only do such models help decide the resources needed to meet an SLA, it also provides guidance to sales personnel on what SLAs to commit to given a set of backend resources.
- K. P. Joshi, Y. Yesha, T. Finin, and Y. Yesha, "A policy-based cloud broker for the VCL platform", Article, International Journal of Cloud Computing, January 2013, 1628 downloads.