Next-Gen MAN: Collapsing the Access, Transport Hierarchy - Technology Information

Telecommunications, June, 2000 by John Jaeger

Bandwidth-intensive services have dramatically changed network topology assumptions within the metropolitan area network.

The ILEC-dominated infrastructure provides TDM circuit aggregation and hubbing across the "feeder loops" and onto the transmission facilities between central offices (COs) has resulted in a strong delineation between access and metropolitan core networks. New dynamics bring this access/metro delineation into question. Customers now demand direct high-speed optical data connectivity to carry mission-critical enterprise applications, Internet traffic, voice and video services. The growth of new communication providers vying to provide industry-leading services on a leased or newly laid fiber plant has dramatically changed the network topology assumptions within the metropolitan area network (MAN). The breadth of protocols and services these carriers must accommodate is increasing, along with the need for rapid service delivery.

With MAN transport requirements moving away from TDM to IP packet-based traffic, facilities-based service providers are taking a hard look at their evolving infrastructure requirements. Operational efficiency is the most significant driving factor in the request for a single optical system platform that can span the range of what a metropolitan network needs. While the motivation for this architectural change is clear, the nature of the optical service node that can satisfy the sheer breadth of these new requirements is not.

Access network provide the direct connection point td the customer, and serve as the demarcation point for man agement and service responsibility between the carrier and customer networks. The access network is responsible for accepting a variety of network protocols and services and adapting them for aggregation and transmission into the core. Requirements have come a long way from the traditional functions of DS0 and DS1 circuit-grooming and the packaging of data services into TDM time slots. The access network must provide distributed switching connectivity for applications riding directly on IP, Ethernet and ATM as well as TDM. Furthermore, optical access speeds have evolved to span the, range from a fractional DS3 service to a dedicated wavelength. A carrier recently commented that, "Access is where my customer is. Delivering an OC-48 or dark lambda service to him means I have to 'distort' the metro collector ring down to that access point." This being the case, the vestiges of the TOM hierarchy are ill suited for today's market environment.

Additional functionality is needed, to aggregate the disparate protocols sup ported and provide the access control and rate shaping they require. This man dates a level of service intelligence to distinguish between application flows and maintain quality-of-service (QoS) information throughout the network. In a significant departure from the past, this service-level discrimination is required at all, levels of the network hierarchy--from the access network ingress point to the differential treatment of an optical dense wavelength division multiplexing (DWDM) wavelength for priority and protection purposes. The requirement to couple the service offering QoS at tributes from the network edge into the optical domain is an additional consider ation for collapsing the access and metro optical core.

The metropolitan core is evolving from the "collector ring" position it currently occupies to one of providing greater flexibility in terms of scaling bandwidth, supporting a range of physical and logical topologies, and providing a more efficient and direct connection into the long-haul optical backbone. Today the metropolitan core is composed of a hierarchy SONET rings, which aggregate access services and transport them to the CO or carrier point of presence (POP) for connection into the long haul network Current functionality is limited to basic transport plus the protection/restoration mechanisms. To address the emerging communication providers' new requirements the metropolitan core needs to scale up to 100 times the capacity of access bandwidth. To provide for the variance across a given metropolitan/regional network and between major metropolitan centers and Tier 2 and Tier 3 cities the metro core solution must cost-effectively scale across a large range of capacities.

An even more difficult problem is sup porting the evolving physical network topologies and associated service traffic patterns Ring and linear add/drop solutions remain appropriate in many environments, but in some cases there are considerable advantages to diversely routed hub-and-spoke and mesh topology support within the MAN. Topology selection depends on.

* Access to and quantity of installed fiber;

* Protection and restoration attributes across all network elements for a given service; and

* Protocols and services to be carried on the infrastructure.

Operational Efficiencies

While there can be considerable savings on equipment costs, the largest factor in the business case for collapsing the access and metropolitan core solution is the array of operational efficiencies that can be realized. Approximately 30 percent of the costs associated with delivering a defined service to a set of customers with a MAN are equipment costs: 70 percent are allocated to service activation and ongoing operations, and management of the network and service.