The paper presents SPs and Enterprise Ethernet/VPLS requirements, network deployment scenarios, the performance and scalability problems of such deployments and the challenges to make the Ethernet Services as cost effective and scalable. It provides different theoretical models, and practical solutions for future VPLS deployments. The methodology developed is top-down, from the Service definition to the network topology, SP infrastructure, and nodal view. The tier approach allows us to better understand how different parameters are influencing the Services behavior, and the overall performance of the network.

Different case-studies, with different types of inter-working (ATM-FR-MPLS, MPLS-Ethernet-ATM/FR), based on our direct deployment experience are presented also.