Despite the gloom and doom of the past year in optical networking, one sector continues to generate excitement and optimism: optical Ethernet.

Just about every vendor with an access system that can claim carrier-grade robustness is staking a claim in the emerging optical Ethernet market. Indeed, labels like optical Ethernet, metro Ethernet, and Gigabit Ethernet are appearing on everything from Sonet ADMs to edge switches/routers to PONs.

Although we're focused on access in this report, it's important to note that metro Ethernet usually involves an end-to-end solution that starts in the metro core and typically involves optical edge systems that also support Ethernet (see A New Optical Taxonomy, page 7, for a list).

Why is Ethernet so popular as a metro solution? It's just too promising a market to ignore. There are a number of reasons for its attraction:

• Most (perhaps 90%) of transported business data now starts and ends as Ethernet frames on LANs.
• The triumph of Ethernet/IP in the business world is essentially complete. Ethernet/IP is the massively dominant corporate level-1/2/3 networking technology: It is cheap, works well, is used heavily, and is widely understood and trusted.
• Ethernet has proved extraordinarily adaptable in the business world, having metamorphosed from its original 10-Mbit/s form, first to 100-Mbit/s Fast Ethernet, then to 1,000-Mbit/s Gigabit Ethernet, and now to 10,000-Mbit/s 10-Gigabit Ethernet.
• So, if Ethernet could be beefed up to carrier-class performance, why use different technologies to carry the Ethernet frames and their IP packet payloads across the metro and into corporate LANs?

Two theoretical approaches. In metro Ethernet right now, there are two basic and opposed views. For simplicity’s sake, Light Reading breaks these into two groups: The first group is the Etherians, those that espouse the idea of Ethernet everywhere. In this scenario, Ethernet appears as the end service to the customer, but in addition to providing that access link, it also furnishes the Layer 2 switching and the base Layer 1 bitstream transport throughout the metro network. The metro not only looks like native Ethernet in terms of what it delivers to the customer, but it is native Ethernet in terms of how it works. Ethernet handles all functions: Layer 2 switching, traffic engineering, resilience, and service provisioning, in all areas of the metro network – aggregation and core as well as access.

The second group is the Soneteers. This group embraces Ethernet over Sonet, in which next-generation Sonet equipment provides Layer 1 bitstream transport and a range of key functions, such as resilience, path management, and bandwidth management. Ethernet does Layer 2 switching, service provisioning, and end services.

In the real world, the Etherian and Soneteer groups are increasingly blending. In terms of access, mix-and-match solutions are appearing, such as ones that support native Gigabit Ethernet for access, emanating from Ethernet over Sonet in the metro core. Bottom line? There is no single metro Ethernet technology or architecture.

Cooperative agreements. Because the problem of delivering Ethernet access in carrier networks typically involves a core and/or edge strategy that includes Ethernet, vendors are ganging up with partners to address the problem. Atrica Inc., for instance, is offering its packet-based edge gear along with access kit from FlexLight Networks (see Atrica Teams With FlexLight).

As time passes, it's likely that all kinds of Ethernet access equipment, including Ethernet-based PONs (see preceding page), will be featured in partnerships with suppliers of transport and edge equipment.

Services in focus. The idea of emerging Ethernet services is to turn the corporate WAN into a big LAN, extending into the last mile – or first mile, depending on one's viewpoint.

A range of service types have been proposed, several of them by the Metro Ethernet Forum (MEF), an industry group that is attempting to define a set of standard service approaches for carrier reference.

Among the service types under discussion at the MEF are the following:

• On-demand granular variable bandwidth services: for instance, 1-Mbit/s to 1-Gbit/s in 1-Mbit/s increments. Proponents say this service would be great for IP access, service provisioning, etc.
• Ethernet Virtual Private Line Services (EVPLS): for instance, point-to-point services replacing current private lines and frame relay links.
• Ethernet Virtual Private LAN Services (EVPLnS): for instance, point-to-multipoint connection of multiple sites as if they were on a single Ethernet Circuit Emulation Service (CES). This kind of service would be suited to transport of TDM services.

Standards in progress. Besides defining different metro Ethernet services, there’s a major effort underway to adapt Ethernet itself to fit carrier networks. Some of this work is being undertaken by the Institute of Electrical and Electronics Engineers Inc. (IEEE), the engineering standards group that defined Ethernet in the first place. As always, the IEEE standards are concerned only with the basic physical and media access control aspects of the technology, and not with higher functions that are specific to the needs of carriers. That's fallen to the MEF and other groups.

Among the IEEE specs being pondered are the following:

• IEEE 802.3z 1-Gigabit Ethernet: This is an established standard for 1,000-Mbit/s Ethernet technology and products. It's used today in dark-fiber metro Ethernet cores and will play an ongoing role in metro Ethernet access.
• IEEE 802.3ae 10-Gigabit Ethernet: This is expected to become a new standard in mid-2002, but products are becoming available now. It's set to play a key role in future metro Ethernet cores, but could also eventually play in access as well.
• IEEE 802.3ah Ethernet in the First Mile: This is expected to become a new standard by middle to end of 2003. It's designed to provide 1-Gbit/s and 10/100-Mbit/s Ethernet in fiber and copper-pair access networks.

As noted, the work of making Ethernet carrier-class requires defining resilience, circuit emulation, guaranteed service levels, and methods of integrating optical Ethernet with Wavelength Division Multiplexing (WDM). What's more, support for operations, administration, maintenance, and provisioning is required. The MEF hopes to unify selected standards from the IEEE and other groups, such as the Internet Engineering Task Force (IETF). It's also intent on adding its own to the mix.

Other efforts are underway too, which may affect metro access in various ways. The IEEE 802.17 draft standard, for example, is aimed at defining the Resilient Packet Ring protocol, which adapts Ethernet to run on metro Sonet rings, preserving the ring resilience. The Resilient Packet Ring Alliance is an industry group that's promoting the adoption of this technique, which has been slow to get off the ground but has the support of Cisco Systems Inc. (Nasdaq: CSCO) and other key vendors.

Added to these standards efforts are proprietary techniques, which are appearing in products now and may or may not make it into various standards, depending on their market reception.

Table 3: Ethernet Access

Related stories in Light Reading:

• Cisco Acquisition Causes RPR Stink
• RPR in the Spotlight
• Quantum Bridge Casts a Wider (Ether) Net
• There's an Ether-Fest at Supercomm

• Atrica Teams With FlexLight
• Nortel Unifies Optical
• Metro Ethernet Forum Grows
• Ethernet Frame Relay
• Hatteras Plans Switch-Hitting Ethernet