Mustapha Aissaoui

Service convergence using MPLS multiservice networks

Enterprises are increasingly using virtual private networks to interconnect remote sites. Traditionally, service providers have used ATM core networks to deliver layer 2 services such as frame relay, ATM, or TDM private lines, which enterprise customers have then used to build their corporate network infrastructure. Such services account for the majority of data service revenues today. However, pressure has increased on service providers to combine increased flexibility with reduced costs in the context of a highly dynamic telecommunications market. Service providers also need to generate new revenues from their IP network infrastructure, through new opportunities such as IP VPNs and virtual private LAN services, while simultaneously achieving operational efficiencies through the convergence of all of their services on a common MPLS backbone. New access and metro network technologies, such as Ethernet, are also emerging that can be used to deliver these new services to enterprise customers alongside ATM and frame relay access. This must be achieved while also supporting existing technologies such as ATM, which continue to deliver highly profitable services. This article discusses the technical challenges in meeting the often conflicting requirements of delivering both traditional layer 2 services and new layer 3 services on a converged MPLS network. We show how both network and service interworking are required, and how these must operate at the user, control, and management planes to enable profitable services to be delivered over the new converged network. The different solutions being defined in the standards bodies are described, and the distinct scenarios they address are explained.

Considerations for scheduling and servicing events using a calendar structure

Networks are supposed to support multiple service categories that include real-time services and best-effort services. In both types of applications, it is important to differentiate between the requirements of different streams, and allow for an appropriate distribution of resources among them. Work-conserving disciplines may be implemented with the use of queue or scheduling calendars. Since calendars have finite resolutions, theoretical emission times that determine the positions of frames within the calendar slots have to be rounded to match the boundaries of the slots. The selection of a rounding method affects the performance of a work-conserving discipline. We present an efficient calendar structure and examine several rounding methods.