Service Level Indication

Abstract

The offering of QoS based communication services has to face several challenges. Among these, the provisioning of an open and formalised framework for the collection and interchange of monitoring and performance data is felt as one of the most important issues to be solved. Indeed, this is true in scenarios where multiple providers are teaming (intentionally or not) for the construction of a complex service to be sold to a final user, like in the case of the creation of a virtual private network spanning multiple network operators and infrastructures. In this case, failures in providing certain required levels in the quality parameters should be dealt with an immediate attribution of responsibility across the different entities involved in the end-to-end provisioning of the service. But also in cases apparently much simpler, for example when an user requires a video streaming service across a single operator network infrastructure, there is a demand for mechanisms for the measurement of the received quality of service across all the elements involved in the service provisioning: the server system, the network infrastructure, the client terminal and the user application. It is clear that this is a complex problem, involving different technologies, disciplines and research areas. In this paper, starting from the ongoing work in the definition of standard interfaces for the Quality of Service negotiation (Service Level Agreements) and control (Service Level Specifications), as well as from the work ongoing in the IPFIX and IPPM working groups from the IETF, we introduce a new document specifically for delivering monitoring information to user applications. We called such a document Service Level Indication. We here aim at sketching a possible starting point for a research discussion.