IP Switching Enhancements over IP Differentiated Services for QoS Interworking

In the last decades, the explosive growth of the Internet gave rise to the development of several telecommunication technologies, aimed at satisfying the increasing users needs with different degrees of Quality of Service (QoS). Together with the development of QoS technologies, an increasing number of Service Providers and networking companies developed their specific architectural and implementation choices. As a consequence, nested, heterogeneous infrastructures compose today's Internet. End-to-end QoS provisioning in such kind of infrastructures requires proper mapping operations among different protocols and united architectures to interconnect different QoS solutions. In this perspective, the paper analyzes QoS interworking architectures. Two solutions are considered: IP DiffServ and IP switching. They are compared with respect to recent results of standardization bodies and of scientific literature.

I. Introduction

The support of end-to-end (e2e) QoS over heterogeneous networks, composed of different portions (also called Autonomous Systems, ASes), is a hot topic of research. The main point is to build an overall e2e architecture that offers full support to QoS, independently of the single solution part of the heterogeneous network. Possible QoS technologies are: ATM, IP Differentiated Services (DiffServ), IP Integrated Services (IntServ), and, more recentely, MPLS. The problem involves a common language for QoS definition, interworking solutions, signalling, and control mechanism implementations. Having in mind tactical hazardous and challenging environments as in military and civil protection world [1], [2], it is often recommendable to have QoS for each user considering Multi Level Priority Pre-emption (MLPP). It means to identify each single e2e connection having a specific Service Level Agreement (SLA).