Abstract
The increasing demand for distributed multimedia applications makes evident the need for end-to-end quality of service (QoS) provisioning. Pmticularly, operating systems, despite their location at end systems, switches or routers, must guarantee that resources under their control are adequately managed to fulfill the application requirements. This work proposes an architecture for adaptive QoS provisioning on network operating systems (QoSOS), focusing mainly on the packet queuing subsystem. The development of such architecture came after an analysis of solutions currently found in the literature and the perception of their functional similarities. QoSOS allows the reuse of common functions and the definition of an internal organization that is equivalent in different systems. In order to demonstrate how QoSOS can be applied in a real QoS provisioning scenario, the paper describes the modeling and implementation of an adaptable Intserv support, focusing on the management of the output queues of the Linux operating system. The architecture instantiation is based on few modifications introduced into the standard Linux kemel, that adds some desirable features such as runtime service adaptation.
Highlights
Quately managed to fulfill the application requirements
This work proposes an architecture for adaptive QoS provisioning on network operating systems (QoSOS), focusing mainly on the packet queuing subsystem
Authors are with the Department of Infonnatics of the Pontifical goal, we present a generic architecture for QoS provisioning Catholic University of Rio de Janeiro, Rio de Janeiro, RJ
Summary
The Service Parameterization framework models the structure that defines the behavior of user flows (load characterization parameters), the QoS requested (QoS characterization parameters), and information about the intemal state of each OS subsystem (like provider petformance, availability, etc.). When a service request is issued, the requesting user must provide the load and QoS characterization parameters of the highest abstraction level of the provided service. Different abstraction levels usuaiiy can for different kinds of descriptions and parameters. The QoS negotiation mechanisms, explained later, are responsible for parameter mappings from a higher abstraction level to lower ones, until those that describe resource behaviors. Whenever a new flow admission request is performed or when the subsystem is a target of some kind of service adaptation, each subsystem's intemal state parameters must change . The association of policies to service categories simplifies QoS provisioning mechanisms
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have