Optical Ethernet offers a logical upgrade route for both enterprise networks and telecommunications carriers.

Ethernet technology based on twisted copper wire underpins much of the corporate network infrastructure in existence today, and many IT managers unsurprisingly see optical Ethernet as a logical upgrade path.

A major reason for this is the recent development of the 10 Gigabit Ethernet standard, which will give network managers bandwidths of 10Gbit/s over fibre-optical cabling, compared with the 1Gbit/s bandwidth currently available. This, plus the continuing growth of e-business on the Internet, make optical Ethernet an attractive proposition for firms and carriers wanting to take maximum advantage and profits from the Web.

Not that optical Ethernet will replace copper-based Ethernet networks immediately or even in the long term. Local area networks (LANs) based on copper wiring will continue to exist in enterprise infrastructures over short distances. However, Ethernet signals degrade over distances of 100m when sent over copper wires, so for wide area networks (WANs) and links between offices, sites or campuses, Optical Ethernet's maximum signal distance of 20km over single mode and 412km over multimode fibre-optical cabling means that it could be a neat high-bandwidth solution that integrates well with existing copper infrastructure.

Linking copper and fibre
The limitations of copper present a problem within many modern buildings and commercial sites. The problem is compounded by the fact that many networks are now beginning to centralise their functions and services. Therefore, central server farms require the type of high-bandwidth that fibre can provide to minimise congestion and delay.

Voice over IP (VoIP) and video conferencing are real-time communication applications that are increasingly demanded by corporates, but these can be only delivered when a total broadband infrastructure is available.

Connecting old and new technologies is easily achieved Ã,­ Ethernet switches that can link copper and fibre are widely available. The network manager simply jacks in the optical cable at one end and the copper cable at the other in order to pass data between the two media.

Suggestions that optical Ethernet components will soon be manufactured in sufficient volumes to drive down their prices seem premature, amid reports of optical cabling shortages, but it may be possible one day to carry optical Ethernet connections right down to the desktop PC as well as the corporate network backbone.

For now, the biggest users of optical Ethernet may be telecommunications carriers providing secure virtual private networking (VPN) links between distributed corporate locations, including branch offices, server farms and remote storage networks. And since copper Ethernet dominates the LAN, the networking community sees the seamless connectivity of optical Ethernet as highly desirable.

Several applications are now emerging that would only offer acceptable quality when large amounts of data can be transferred in real time Ã,­ electronic whiteboarding and video conferencing, for example.

Telco carriers can support this type of service through regional optical Ethernet networks extending high-speed access to remote locations with virtually no degradation in service.

The metropolitan area network (MAN) access link is the last mile between the telecoms carrier and the corporate premises. The network edge Ã,­ the point linking the carrier to the customer Ã,­ can be located at the customer premises interface or it can be located the carrier's point of presence (POP).

This access link has often in the past been based on copper-wire technologies such as time division multiplexing (TDM), and has been relatively slow. Fibre deployment in major MANs and advances in last-mile access technologies, such as wireless broadband and optics, remove this as a bottleneck and permit Ethernet to be used as the only protocol.

Data traffic volume has already surpassed voice traffic volume in many MANs. The Synchronous Optical Network/Synchronous Digital Hierarchy (Sonet/SDH) transport infrastructure, typically based on a 155Mbit/s ATM architecture, is not optimised for data traffic and cannot scale to support the rapid growth of data. Carriers are looking for cost-efficient, data-optimised solutions to replace their existing Sonet/SDH infrastructures, and in many cases have settled on optical Ethernet as the answer.

Service creation is currently in vogue in the carrier industry, with telecoms firms trying to create more applications that they can sell to their corporate customers. Bandwidth control and rapid service provisioning are among the important functions that network equipment must now perform. Managing data flows, multiplexing lower-speed data streams, and limiting network accessibility are examples of functions that manage the consumption of network bandwidth. The ability to deliver services to customers at short notice can be crucial.

Interoperabilty options

As optical Ethernet networks begin to deliver huge bandwidth to end users, network managers will begin to demand differentiated services. Various functions can be applied to the incoming traffic and to traffic flows to improve performance.

While building new optical Ethernet infrastructures is a logical step for startup carriers, the incumbent carriers will want to accommodate Ethernet without totally redesigning their local networks.

As more and more networks migrate to IP, interoperability between IP and other communications technologies will become a larger concern for carriers. Support will be available for all copper, cable, and optical fibre media types for carriers and corporates. All the major forms of networking infrastructures will soon be supported, including ATM, Packet over Sonet, dense wave division multiplexing (DWDM), Gigabit Ethernet, T1 and T3, and coaxial cable.

Carrier networks will be developed in a way that guarantees interoperability with the vast base of existing LANs and WANs. But carriers are unlikely to replace existing ATM switches with optical Ethernet equivalents just to please corporate customers, because the cost/benefit calculations are unlikely to justify such moves.

There are several options available for upgrading carriers' networks to optical Ethernet. One involves IP packets being transported directly over a full Sonet/SDH infrastructure using Point-to-Point Protocol (PPP) encapsulation or over an Ethernet segment built from Sonet/SDH links. Sonet/SDH would be preferred for long distances and could be also be used at the MAN level. The problem here is that full Sonet TDM capabilities, the Sonet physical layer, and the Sonet management functions would all be included, resulting in considerable bandwidth overhead.

Another option is packet over Ethernet over Sonet/SDH (PES). This is a method based on technologies that are well understood. Under this approach, Ethernet is scalable from 10Mbit/s up to 10Gbit/s, allowing the bandwidth offered to be more closely matched to the anticipated demand.

This configuration adds a switching capability that can reduce the number of point-to-point links, so a full mesh may not be required. Ethernet therefore allows LAN, MAN, and WAN networks to be combined to form end-to-end connections, thereby reducing the need for format and protocol conversions.

Packet over Ethernet over WDM (PEW) is a slight variation on PES and is used to transport Ethernet over a wave division multiplexed physical layer. A key advantage of PEW is that the bandwidth overheads of ATM and Sonet/SDH can be eliminated. Both 1Gbit/s Ethernet and 10Gbit/s Ethernet are more affordable, more practical, and simpler than ATM.

A problem that may hinder the adoption of optical Ethernet by carriers is that the technology may be more awkward for carrier network managers than for corporate network managers. This is because carriers have used ATM-based Sonet/SDH architecture, so they have no existing legacy of Ethernet skills. These carriers will not take their ATM switches out and install optical Ethernet switches instead, unless they are presented with sufficient financial incentives from customers to do so. Given the declining value of bandwidth, this seems unlikely.

Copyright © 2009 CBS Interactive, a CBS Company. All Rights Reserved.
ZDNET is a registered service mark of CBS Interactive. ZDNET Logo is a service mark of CBS Interactive.