Carrier Ethernet | Teleinf Edu

From LAN to WAN: Extending Ethernet Services

Ethernet was developed for Local Area Networks (LANs). Simple operation, low cost, and successive speed increases supported its use in buildings and campuses. Connecting offices in different cities, across a Metropolitan Area Network (MAN) or Wide Area Network (WAN), historically required carrier services based on different technologies and interfaces. The customer received a link termination at each site, while the carrier managed transport between those points.

Earlier WAN services included:

A private, permanent digital connection between two locations. It offered predictable operation, but limited capacity and high cost. The rate belonged to a specific circuit and did not scale smoothly with changing demand.
A flexible packet-switched service that required its own terminology and configuration; circuit capacity could be oversubscribed. Service parameters were specified separately for virtual circuits and their agreed rates.
A cell-switching technology for carrying voice, video, and data on one network, with complex and costly deployment requirements.
Synchronous optical transport systems based on TDM. They initially focused on telephony, while also carrying other fixed-rate signals.

Internal networks used Ethernet frames, while interconnecting sites often required mapping them to a carrier WAN service. That added interface hardware, different configuration requirements, and specialist knowledge. Carrier Ethernet addresses the need to carry Ethernet services between distant locations as well. The customer keeps an Ethernet service interface, while the provider core may use other transport technologies.

Carrier Ethernet and Its Five Core Attributes

Carrier Ethernet (also known as Metro Ethernet) is not a separate frame format. It is a set of servicesthat extends Ethernet from the LAN into a carrier's MAN or WAN, with defined requirements for reliability, scalability, and service management. MEF definitions cover the service type, access points, traffic parameters, and monitoring method.

The MEF (Metro Ethernet Forum) defines five core attributes for these services:

Standardized Services: The customer interface uses an Ethernet service format defined by MEF and IEEE. This simplifies service design and management without requiring the customer to convert Ethernet to another protocol. The definitions also describe UNIs, EVCs, and service attributes.
Scalability: A service can scale in bandwidth and in the number of locations. For example, a carrier may increase a connection from 100 Mbps to 1 Gbps or add an office without redesigning the customer network. Available bandwidth steps depend on the service and provider infrastructure.
Reliability: A service may use fault detection and protection mechanisms, while OAM can verify continuity. In specified configurations, protection switching can target restoration within 50 milliseconds; this objective comes from transport-network requirements such as those associated with SDH/SONET and is not a guarantee for every service.
Quality of Service (QoS): Carrier Ethernet can provide agreed performance parameters instead of relying only on public Internet best effort. This is expressed through a that specifies metrics such as bandwidth, availability, latency (delay), jitter (delay variation), and frame loss.
Service Management: The carrier provides tools for provisioning, monitoring performance, diagnosing faults, and checking SLA parameters. These functions are often referred to as OAM (Operations, Administration, and Maintenance). They allow the provider to test continuity and report service state at defined maintenance points. The resulting records support fault isolation and SLA review across the provider domain.

The Carrier Ethernet Service Types

MEF defines service types describing connections between customer locations. Their basic topologies are point-to-point, multipoint-to-multipoint, and rooted multipoint.

E-Line (Point-to-Point)

An E-Line service creates a point-to-point connection between two locations. It provides a leased-line-like service through an Ethernet interface and Ethernet service mechanisms. MEF service definitions include port-based and VLAN-based variants.

Example: the service connects an office in San Francisco with a data center in Las Vegas as one point-to-point connection.

E-LAN (Multipoint-to-Multipoint)

An E-LAN service connects multiple locations and allows any-to-any communication. It provides a Layer 2 service across a wide area, so connected sites can use a common forwarding and addressing model. One service can have multiple access points instead of a separate circuit for every pair of sites.

Example: regional offices are connected so they can share files and access servers without separate point-to-point services.

E-Tree (Rooted-Multipoint)

An E-Tree service creates a rooted multipoint topology. A central root site can communicate with all leaf sites. Leaf sites can communicate with the root, but they cannot communicate directly with one another. This model suits content delivery and centralized services where branch-to-branch traffic is not allowed. The root and leaf roles are service attributes, so the forwarding policy is defined independently of the physical cable layout.

Example: a central system distributes content to every leaf site, while leaf sites do not send it directly to one another.

Technologies Used in Carrier Ethernet

Carrier Ethernet services are delivered through a provider's core network using technologies that separate services, scale capacity, and protect traffic.

Service VLAN tagging (Q-in-Q):A customer LAN may use VLAN tags defined by IEEE 802.1Q. A provider can add an outer service tag using Provider Bridge mechanisms, historically specified in IEEE 802.1ad. The S-tag identifies the service and lets provider bridges separate customers and support identical inner C-VLAN identifiers for different customers. The customer tag can remain present across the provider service, while forwarding in the provider network uses the S-VLAN.
:MPLS assigns packets to forwarding classes at the network edge and carries labels in a stack. In the core, Label Switching Routers use the label to select the next hop and an operation on the stack without reexamining the inner network-layer header. Label Switched Paths (LSPs) support Traffic Engineering, which directs traffic along a selected path to limit congestion, as well as Layer 2 and Layer 3 VPNs. The ingress assigns labels from the information available at the edge, while the egress removes them or delivers the packet to the appropriate service.
Carrier-grade transport layers (OTN):Carrier Ethernet can be a client signal for an Optical Transport Network. ITU-T G.709 specifies OTN framing and client mapping, management functions, and forward error correction (FEC). Depending on the variant, Ethernet signals such as 10, 100, or 400 Gbps can be mapped and transported over a WDM backbone. OTN adds monitoring and correction to optical transport without changing the Ethernet service exposed at the customer interface.
Protection and resiliency protocols:ITU-T G.8032, also known as Ethernet Ring Protection Switching (ERPS), defines automatic protection switching for ring topologies. One link is blocked for traffic to prevent a loop. When a failure is detected, the R-APS mechanism changes the ring state, updates forwarding, and makes the standby path available; specified configurations target restoration within 50ms.