Comparing SLAs for Cloud Services: A Model for Reasoning

Abstract

Nowadays cloud services are gaining their momentum. A Service Level Agreement (SLA) represents an agreement between a service provider and a customer for a particular service provision. Cloud providers and services are often selected more dynamically than in traditional IT services. Hence, services need to be compared according both to technical aspects and the promised SLAs, but no widely accepted model, standard or best practice, that would lead to a more rigorous SLA comparison, are currently available. In this paper, we present a conceptualization of SLAs including the explicit modeling of formulas for service level indicators, and its corresponding open, reusable ontological representation. On its top, we provide operators for analysis and comparison of SLAs, relying on a reasoning framework in Logic Programming capable to manipulate indicator formulas and other SLA features. We provide an evaluation by considering a cloud scenario where several SLAs have to be compared.

SLA-aware operational efficiency in AI-enabled service chains: challenges ahead

“…In other words, the SLA terms are highly specific to both customer and the type of market in this example, and their definitions most likely semantically differ from typical generic operational KPIs used by e.g., an IT outsourcing provider. Another example for similar, but semantically different SLIs for service availability is given in Longo et al (2018) and reproduced in Table 1. Further examples are given in the scenarios depicted in Sect.…”

“…While the problem of data availability in a broader sense has been extensively studied in the current literature on schema matching and ontology matching (e.g., Shvaiko and Euzenat 2005;Noy 2004;Bernstein et al 2011), the more specific problem of mapping SLIs to KPIs in a finegranular manner and including the ability to reason on the differences between intermediate computation results in different stages of aggregation has been largely overlooked so far, except for some works in early stages (cf. Longo et al 2018) or in the context of specific use cases (e.g., Bellini et al 2018).…”

“…Similar approaches may prove useful for mapping SLIs to KPIs, and vice versa. • Regarding semi-automated approaches, prior works have focused on modeling SLAs in OWL and applying Logic Programming (i.e., Prolog) to reason about different semantic definitions of SLIs among different SLAs with the goal of facilitating their comparison (Longo et al 2018). Such approaches may be potentially extended to inform and aid semi-automated methods for mapping SLIs to KPIs.…”

A Systematic Literature Study on Definition and Modeling of Service-Level Agreements for Cloud Services in IoT

“…Service Level Objective: A Service Level Objective represents a threshold or a value on an SLI or a metric. Basically, an SLO depicts the commitment of the service provider to the consumer on a particular service [14], [15].…”