Service Quality Model Evaluation

Service Quality Modeling

In the beginning of 2000 a term has been coined perhaps more than any other in relation to Quality of Service (QoS) and Service-Oriented Computing (SOC), that is, of the Service Level Agreement (SLA) (Tian, Gramm, Naumowicz, Ritter, & Schiller, 2003), (Sahai, Machiraju, Sayal, Jin, & Casati, 2002), (Keller & Ludwig, 2002). Of course, the term itself had been defined a long time ago as “… an agreement between the provider of a service and its customers which quantifies the minimum quality of service which meets the business need” (Hiles, 1994). In short, SLAs are the formalization of a provable, quantitative assessment of QoS upon which both the provider and the consumer negotiate and agree. Upon agreement the provider must commit the necessary resources so as to satisfy the consumer requirements. The consumer is then able to evaluate the outcome of the service against the set of SLA terms. There are various negotiation models that differentiate this procedure. The main differentiation axes (Figure 1) are the lifetime of the SLAs and the granularity of the SLA terms. The differentiation mainly affects the side (provider or customer) that has to do the evaluation and in turn the side which is more susceptible to SLA violations (normally these sides are opposites).

Figure 1.

Variety in SLA models. The further right we move on the “SLA lifetime” axis, the more a provider has to reserve its resources for the customer. The further up we move on the “granularity” axis, the closer we get to the use of terms that match the customer's understanding