In the late 1990s, many metro service providers based their business plans on the assumption that bandwidth limitations on metropolitan-area networks (MANs) were the major bottleneck to higher-capacity long-haul networks. The construction spree that followed replaced the metro bottleneck with a considerable metro fiber surplus that, despite many consolidations, persisted through 2006.
The hope today is that with the right product portfolios, local access and metro service providers can capitalize on burgeoning demand for enterprise services to put this surplus to use. With growing capacity requirements, strong demand for traffic differentiation and prioritization, and a movement toward the converged WAN, delivering these more complex services will result in less customer churn and growing bandwidth requirements.
National and multinational enterprises employ a vast array of tools to manage their internal and external communication and data storage requirements. Increasingly complex WANs now integrate and carry voice, video, real-time and business critical data, and low-priority data streams. Typical service offerings are complemented with access to the public Internet and, frequently, remote storage capacity. With business continuity a primary focus, the redundancy and restoration capabilities of the service provider become paramount. End-to-end service-level agreements that guarantee near 100% availability, specify stringent latency and jitter requirements, and dictate diverse physical routing are becoming the norm.
WANs deliver end-to-end connectivity between customer sites. Providers depend not only on their core networks, but also on partnerships with local access carriers. Whether the connectivity offering is a LAN extension service in a single city or a secure MPLS IP VPN across 15 international capitals, each requires physical access to the enterprise facility and an interconnection with a central carrier network to aggregate and distribute traffic. MANs and local access connectivity are significant cost components of the total network service.
The distinction between MAN providers and local carriers is somewhat ambiguous. Some of the largest metro bandwidth providers are themselves incumbent local exchange carriers (ILECs) while some of their MAN-oriented competitors provide direct fiber connectivity to thousands of end-user buildings. Rather than specializing in last-mile connections, however, the primary function of metro networks is linking access and long-distance networks. Thus, the buildings MANs primarily connect are traffic aggregation centers-the points of presence (PoPs) where all three network types meet.
Though not all MAN operators sell to end users, MAN owners offer break-out points reasonably close to end users, especially large commercial premises. The MAN provider or another carrier can connect a “local tail,” joining the MAN ring and the end user. Thus, in addition to linking traffic aggregation centers, most MAN operators also deploy rings around the most important commercial business districts in a city, especially areas with high concentrations of high-tech and financial services companies.
For many service providers, the high cost of this end-user access can be the one component that makes their enterprise network bids less competitive. The total cost of ownership (TCO) of multisite network IP VPN offerings includes several network components: port charge at the provider edge, customer site equipment rentals, management fees, class of service charges, and local access. TeleGeography’s Enterprise Pricing Service reviewed sample RFPs for multisite IP VPNs in the third quarter of 2006 and found that local loops accounted for a significant portion of the TCO across global regions. Local access composed between 20% and 42% of TCO in the US, a median of 43% in Europe, and a median of 16% in Asia.
Metro bandwidth and fiber providers see two related enterprise market issues: the low price for corporate WAN services is driving last-mile bandwidth demand to higher levels; and network providers are pushing for lower local access costs, since margins on the network level are thin. Service providers were dependent on expensive connectivity, and because there was significant competition they discounted their core networks, reducing MPLS IP VPN port pricing (the network cost of the solution).
Increasingly, IP VPN service providers are looking for alternatives to traditional leased lines for local tails. In Europe, DSL is replacing the local loop connections of 2 Mbits/sec and below at a fraction of the cost. In the US and parts of Asia, Ethernet connections are becoming increasingly common. Ethernet is particularly enticing because it assures that each customer site has room for growth and that capacity use can be flexible. Upgrading a site from 4 to 6 Mbits/sec on an Ethernet circuit does not require the acquisition of another leased line. Equipment costs of Ethernet services are thus substantially lower for the customer than legacy technologies.
These factors have resulted in significant pressure on local access costs. For the local access/metro sector the install base of sub-T1 and sub-E1 circuits will erode and be replaced by less expensive, more cost-effective, and more flexible Ethernet and DSL access. This is a critical win for service providers delivering managed services. If they can realize cost savings at the edge, they can preserve their core network prices for a longer period.
Another critical issue from a metro service provider’s perspective is dark fiber sales. Customers often realize that purchasing, lighting, and managing their own metro fiber rings themselves will typically be cheaper in the long run because added circuits come at little incremental cost. The service provider selling dark fiber will likely forgo any future lit revenues from the customer, at least for that specific connection.
Because of the fiber glut in most major cities, however, MAN providers are not reluctant to sell dark fiber. For its enhanced Metropolitan Area Networks Database (MANs), TeleGeography surveyed dark fiber infrastructure providers in major business capitals worldwide to quantify supply and to identify specific sites with the greatest quantities of dark fiber available. The speculative builds of the late 1990s left carriers with hundreds of fiber pairs dormant. These companies recognize that the fiber they have deployed is, in many cases, a sunk cost. This resource has a shelf life, and they need to sell it.
The outlook for bandwidth and fiber providers that deliver end-user connectivity or, at a minimum, extend the provider edge as near to the end user’s physical site as possible is mixed. Clearly demand will grow, but if they are not offering the most efficient, cost-effective, scalable services, they may find themselves left out. ILECs and CLECs will increasingly have to review their product portfolios, adapting to an environment where 2-Mbit/sec DSL links can be as little as 10% of the cost of a leased line that delivers 512 kbits/sec. Sub T1 and E1 leased-line local access circuits will increasingly become undesirable.
Dark fiber will likely be offered as gateways to out-of-region services. With the ease and lack of expense with which Ethernet can be deployed over fiber, metro providers should encourage fiber sales, as this will lead to more off-net traffic that they can charge for. Enterprises will likely look to dark fiber when they have sufficient internal traffic demand to make bringing operations in house a breakeven proposition. With sufficient expertise it is thus likely that they can cap operational costs while accommodating growth in demand.
Rob Schult is a research director at Telegeography (www.telegeography.com).