From LAN to WAN: The Challenge of Extending Ethernet

Ethernet, as we know it, was born to dominate the Local Area Network (LAN). Its simplicity, low cost, and ever-increasing speeds made it the undisputed king of networking within a single building or campus. For decades, when a business needed to connect offices in different cities (in a Metropolitan Area Network, MAN, or Wide Area Network, WAN) they had to turn to different, more complex, and significantly more expensive technologies offered by telecommunications carriers.

This "WAN world" was historically built on technologies like:

• Leased Lines (e.g., T1/E1): Essentially a private, permanent digital phone line between two locations. Reliable but very expensive and offered low bandwidth.
• Frame Relay: A more flexible packet-switched service, but it came with its own complex terminology and configuration, and bandwidth was often oversubscribed.
• ATM (Asynchronous Transfer Mode): A powerful technology that promised to unify all network traffic, but was incredibly complex and expensive.
• SDH/SONET: The reliable, synchronous optical backbone of the telecom world, designed primarily for voice traffic in fixed-size circuits.

Businesses faced a significant challenge: their internal networks (LANs) were simple, fast, and spoke the language of Ethernet. But to connect their LANs together, they had to translate this simple protocol into the complex and rigid language of the WAN carriers. This required expensive interface hardware and specialized knowledge. The question arose: "Why can't we just use Ethernet everywhere?"

What is Carrier Ethernet? The Five Core Attributes

Carrier Ethernet (also known as Metro Ethernet) is the answer to that question. It is not a new type of Ethernet but rather a set of services that extends Ethernet from the LAN into the carrier's network (MAN/WAN). It combines the simplicity and cost-effectiveness of Ethernet with the reliability, scalability, and management features that carriers and businesses demand.

The MEF (Metro Ethernet Forum), an industry consortium, standardized what it means to offer "Carrier Ethernet." A service can only be called Carrier Ethernet if it meets five core attributes:

1. Standardized Services: Carrier Ethernet uses the same, familiar Ethernet frame format. This simplifies network design and management, eliminating the need for complex protocol conversions. It provides a standard, technology-agnostic interface to the customer.
2. Scalability: Services must be scalable in terms of bandwidth and the number of locations. A business should be able to easily increase its connection speed (e.g., from 100 Mbps to 1 Gbps) or add a new office to their network without a major overhaul. Bandwidth can typically be adjusted in fine increments.
3. Reliability: The network must meet the stringent demands of business-critical applications. This means the service must be highly resilient to failures. Carrier Ethernet networks incorporate mechanisms for rapid fault detection and traffic restoration, often within 50 milliseconds, a standard inherited from traditional telecom networks like SDH/SONET.
4. Quality of Service (QoS): Unlike the "best-effort" nature of the public internet, Carrier Ethernet must provide predictable performance. This is achieved through a Service Level Agreement (SLA) that guarantees key performance metrics like bandwidth, availability, latency (delay), jitter (delay variation), and frame loss.
5. Service Management: The carrier must provide a comprehensive set of tools for managing the service, including provisioning (setting up the service), monitoring performance, troubleshooting issues, and ensuring the SLA is being met. This is often referred to as OAM (Operations, Administration, and Maintenance).

The Carrier Ethernet Service Types

The MEF has defined a standard set of service types that describe how customer locations are connected. These services are defined by their topology.

Technologies Powering Carrier Ethernet

Carrier Ethernet services are delivered over a service provider's core network using a variety of sophisticated technologies to provide the required isolation, scalability, and reliability.

• Advanced VLAN Tagging (Q-in-Q):While a standard business LAN uses VLAN tags (IEEE 802.1Q) to separate its own departments, a service provider needs to separate entire customers from each other. To do this, they use a technique called Q-in-Q (or IEEE 802.1ad). The provider wraps the customer's already-tagged Ethernet frame inside another, outer "service tag." This outer tag is used to route the frame across the provider's network, keeping all of one customer's traffic completely isolated from another's, even if both customers are using the same internal VLAN IDs.
• MPLS (Multi-Protocol Label Switching):MPLS is a cornerstone of modern carrier networks. It creates "superhighways" or tunnels (called Label Switched Paths - LSPs) across the provider's core network. When a customer's Ethernet frame enters the MPLS network, it is placed into a "shipping container" with a simple label. Routers inside the network only need to look at this short label to know where to send the packet next, which is much faster than doing a full IP routing lookup. MPLS is what enables carriers to perform sophisticated Traffic Engineering (directing traffic along specific, non-default paths to avoid congestion) and create robust Layer 2 and Layer 3 VPNs.
• Carrier-Grade Transport Layers (OTN):Often, the Carrier Ethernet service itself is a "client" that is transported over an even larger, more powerful optical network. The Optical Transport Network (OTN) is a modern protocol suite (a successor to SDH/SONET) that acts like a digital wrapper for high-speed signals like 10, 100, or 400 Gbps Ethernet. It provides robust error correction (FEC) and management capabilities, allowing Ethernet services to be reliably transported across continents over a WDM backbone.
• Protection and Resiliency Protocols:To meet the 50ms restoration goal, carriers use protocols like G.8032, also known as Ethernet Ring Protection Switching (ERPS). This protocol allows switches to be connected in a physical ring topology for redundancy. ERPS ensures one link in the ring is logically blocked to prevent loops. If any other link in the ring fails, the protocol can unblock the standby link in under 50ms to restore connectivity.