Solution Benefits
Metro Ethernet has long been a foundational infrastructure that delivers Layer 2 Ethernet business, federal, and residential services. Carrier Ethernet, largely defined by the Metro Ethernet Forum, establishes the transport and services framework within the Metro Area Network (MAN). The traditional characteristics of Metro Ethernet Services include Layer 2 connectivity models, which support point-to-point, point-to-multipoint, and multipoint-to-multipoint service types. End-to-end layer 3 business access is typically facilitated by L3VPN and extended for high-speed Internet access. In this JVD, we further extend EVPN services to include Internet access. In addition, the behavioral aspects are standardized by the MEF and include service assurance mechanisms, such as E-OAM and Quality of Service (QoS) constructs. The modern metro network has evolved to support a highly capable architecture and more sophisticated feature-set driven by the cloudification of the metro, and the emergence of new complexes like edge compute and telco cloud. Cloud services, applications, and new use cases place increased demands and challenges on the network.
The scope of the Metro Ethernet Business Services (EBS) JVD seeks to address the traditional L2 business access and dedicated Internet access services while also incorporating modern service delivery protocol sets, including EVPN-VPWS, EVPN-FXC, EVPN-ETREE, and EVPN-ELAN with high availability. We tackle the connectivity challenges that are introduced with cloud metro solutions by providing the service connectivity models that are required for interconnection with cloud edge infrastructures and parallel Layer 3 access. In addition, we explore the integration of traditional VPN services like L2VPN, VPLS, and L2Circuit for business and wholesale use cases and the interconnection of these services with the cloud metro architecture.
The topology focuses on the Juniper Cloud Metro portfolio, including the ACX7000 series and MX304 multiservice edge routers as primary devices under testing (DUTs) and PTX10001-36MR routers for core and peering roles with additional helper nodes in the access regions that include the ACX710, ACX5448, and MX204 platforms.
The reference architecture deploys an infrastructure designed to support traditional metro access ring topologies with lean edge services termination. In addition, the topology features a two-stage metro fabric spine-and-leaf design with border leaf nodes performing the lean edge role and facilitating connectivity into edge cloud complexes. Both infrastructures support seamless interconnectivity within and between different access regions. We build the Cloud Metro infrastructure using spine-leaf fabric and multi-ring architecture that facilitates x-to-anything connectivity models and leverages seamless Segment Routing with fast failover TI-LFA recovery mechanisms. Multi-instance ISIS enables the partitioning of the network domain into independent IGP instances to improve scale and contain blast radius. Flexible Algorithm with Application Specific Link Attributes (ASLA) enable the creation of additional layers of abstraction that form distinct paths through the network based on delay or traffic engineering metrics. Transport classes and BGP Classful Transport (BGP-CT) enable the mapping of services onto color transport for both intra-AS and inter-AS services. Traffic is steered through the network based on the defined service-level objectives (SLOs). Flex-Algo Prefix Metrics (FAPM) enable inter-domain traffic steering across flex-algo IGP boundaries and cascade resolution schemes support the transition of services between performance hierarchies during failure events. BGP Labeled Unicast supports the coexistence of inter-AS services with color-mapped services and seamless failover from colored paths onto uncolored paths if required.
