While fiber connected buildings are far from the norm when considering the total universe of commercial/office MTUs, in the world's prominent Central Business Districts (New York, Tokyo, London, etc.) fiber metro rings have been built that encompass a high percentage of city's buildings, bringing them into close proximity to the fiber network. A breed of metro area network provider has emerged to capitalize on this wealth of fiber and deliver very high-speed connectivity to MTUs in these Central Business Districts (CBDs).

The heart of the fiber MAN (Metro Area Network) is the core metro ring, the portion of the network that carries the most traffic and runs through the CBD. The metro core functions as the gateway to the larger WAN. Metro access rings, often referred to as the last mile, are built onto the metro core. The metro access rings are considerably smaller in geographic scale as compared to the larger metro core, in the magnitude of five miles in circumference versus 50 miles in circumference. Short distance fiber laterals are built from the metro access ring to connect nearby MTUs. Source: In-Stat/MDR

While traditional fiber implementations have employed SONET, Ethernet has emerged as a competitive alternative. The benefits of Ethernet in the MAN are principally two-fold, cost and IP service delivery. With lower network component costs, it is thought that providers will be able to pass the cost savings on to customers through lower service fees. For example, Cogent Communications offers a 100 Mbps connection over an Ethernet network priced at $1,000 per month. Compared to the standard T-1 pricing for SONET based access at nearly $1,000, the Ethernet offering posses significant cost savings to the subscriber.

An additional benefit of Ethernet MANs is the flexibility and breadth of service offerings possible. Unlike a SONET-based service, which imposes rigid bandwidth allotments, Ethernet services offer the flexibility to deliver any amount of bandwidth the customer may require, should the provider wish to offer this option to their subscribers. Furthermore, service levels can be dynamically upgraded or degraded almost instantaneously, allowing subscribers to custom fit their bandwidth according to their changing needs. A final benefit of Ethernet is the simplicity of an all IP network, avoiding the complexities of IP/ATM conversions present in many other network topologies.

While the impetus behind Ethernet MANs is building with the emergence of the Ethernet in the First Mile (EFM) initiative and the backing of industry leading equipment manufacturers, there are several issues that are slowing the conversion to all Ethernet networks. From the providers' perspective, aside from the next generation providers such as Cogent Communications, Fiber City, Yipes and others, the legacy providers will be resistant to restructuring their networks. Given the significant presence of SONET equipment in existing networks, it will be difficult to prove a business model to the legacy providers that warrants the conversion to Ethernet MANs.

Additionally, there are quality concerns for Ethernet-based networks, the principal issue being network restoration. While SONET is capable of supporting a 50 millisecond (ms) restoration, the standard imposed to support legacy Time Division Multiplex (TDM)/voice networks, Ethernet-based networks often measures restoration in minutes. In addition to the transport issues, there is also concern with the switching elements for voice traffic. In a SONET network, these switches have a delay in the hundredths of microseconds, while Ethernet switches operate in microseconds. The slower delay in Ethernet switches could induce echo cancellation and jitter.

A final issue that may impede Ethernet MAN deployments is the immaturity of the advanced IP service delivery market. The market for IP voice, video, and other Value Added Services (VAS) are growing, however the market, even as a percentage of revenue for the next generation providers, is still relatively small. Providers are just beginning to truly introduce independently provisioned high margin value added services to this market. As a new service offering, layered or value added services still comprise only 5 to 10 percent of total revenue for these providers, while 90 percent plus of revenues are generated from the broadband connection. Traditional providers, from both the telco and cable camps, are even further from building a business model that is weighted toward IP service delivery and that would thus justify the cost of network conversion.

While absolute conversion to Ethernet in the MAN is several years off, the leading edge of this trend will first impact the MTU environment. Providers in this space will target the most prominent Class A, high profile, CBD buildings for these deployments, as these are closest to the core fiber ring and offer the highest return on investment. The MTU will offer these providers a high-touch environment with high subscriber density, to market their services. The dynamics afforded by the MTU environment will best use the resources that these providers have poured into their networks.