Negotiating those metro-network hazards
Metro Solutions / SPECIAL REPORTS
Tips for designing, selecting, and deploying a metropolitan DWDM network.
SHAHIRA RAINERI and DR. HUBERT JAEGER, Optisphere Networks Inc.
Prompted by competitive pressures in their own markets, business customers across all industry sectors depend more and more on Internet-based applications and the network services that support those applications. Typically, end users are not interested in the specifics of underlying technologies. Rather, they simply want a broad range of affordable, high-speed bandwidth services that can be tailored to their individual requirements.
For service providers operating in the intensely competitive telecommunications market, this growing diversity of customer requirements defines the criteria for today's metro-area networks. Specifically, service providers must build metro infrastructures that are:
- Scalable, allowing them to expand, quickly and affordably, the number of customers, customer locations, and end users on the network.
- Flexible, so they or their customers can provision bandwidth dynamically, when and where individual customers need it;
- Fast, with speeds ranging from DS-1 (1.5 Mbits/sec) to OC-48 (2.5 Gbits/sec) and higher, so the network can support a wide variety of applications, including business-to-business transactions, videoconferencing, and real-time database mirroring.
- Transparent, to accommodate multiple data formats and protocols such as Internet Protocol (IP), ATM, Gigabit Ethernet, SONET, Escon, and serial digital interface (SDI).
Although laying more fiber is one solution to the metro-networking challenge, it is not a realistic one for most service providers. Even assuming the necessary rights-of-way are available, few service providers operating in today's competitive market can afford the time and money required to deploy additional fiber. Consequently, service providers are looking for technology-based solutions that enable them to satisfy customer needs, yet still get the maximum capacity and value from their existing metro-area fiber investments.
Basically, they have three choices: ATM, time-division multiplexing (TDM), and DWDM. Certainly, ATM and TDM each can deliver multiple services from a single platform. However, both are limited in terms of speed and scalability. They cannot readily support high-bit-rate services such as Gigabit Ethernet, Escon, and SDI. Further, the relative rigidity of ATM and TDM technologies means service providers can expand their operations only through forklift upgrades, processes that are far too slow, cumbersome, and expensive for today's marketplace.
Not surprisingly, more and more service providers are beginning to implement DWDM-based metro solutions. Featuring speed, scalability, manageability, and protected reliability, DWDM technology brings to the metro marketplace the same cost and networking efficiencies that it already provides in the long-haul arena. DWDM solutions designed for use in city networks typically feature transmission links of up to 100 km. Many have a modular architecture that enables service providers to scale them quickly and cost-effectively, for example, from one to 32 channels per fiber (see Figure 1). In response to customer appetites for speed, some metro DWDM solutions also support maximum bit rates per wavelength, ranging from 50 Mbits/sec all the way up to 10 Gbits/sec.
The diversity of service providers operating in the metro marketplace, as well as the differences in their individual business strategies and customer bases, calls for metro DWDM solutions that can be tailored to an individual carrier's requirements. In fact, the ideal metro DWDM solution must be sufficiently flexible to allow a given service provider to adapt its particular business strategy to changing market conditions-and satisfy a range of speed and capacity needs across its customer base.
This type of adaptability is particularly important because the implementation challenges facing an incumbent service provider differ from those confronting a new market entrant. For one thing, most established service providers have substantial investments in legacy SONET/SDH infrastructures, which they typically deploy in point-to-point and/or ring configurations. They obviously want to protect those investments while, at the same time, gradually evolving their metro networks to a meshed DWDM infrastructure.
Incumbents, therefore, want a pay-as-you-grow solution that enables them to partition their network and achieve maximum return on their SONET/SDH investments, while also offering them sufficient increments of DWDM capacity on a cost-justified basis. Given customers' increasing appetite for speed, service providers are well aware that "sufficient" DWDM capacity means futureproof DWDM capacity.
As recently as 12-18 months ago, a 16-wavelength metro solution probably was adequate to satisfy most service providers' needs. Today, however, more and more customers want Gigabit Ethernet or 10-Gigabit Ethernet, with some even looking ahead to 40-Gbit/sec solutions. Understandably, cost-conscious service providers want to implement a scalable metro DWDM solution that will not be exhausted in six months to a year, neither in terms of the number of wavelengths nor the interface speeds.
Another implementation challenge facing incumbent providers centers on the network-management aspects of a metro DWDM solution. Many business customers today want more than a portfolio of discrete services. Coping with their own operational, budgetary, and time-to-market pressures, they want to be able to control at least some of the network services they obtain from their providers.
Consequently, incumbent providers are looking for a metro DWDM solution they can embed in their existing operational support systems (OSSs). Such a solution enables their OSSs to handle tasks such as order management, provisioning, and trouble ticketing, while allowing their customers to provision their own bandwidth on an as-needed basis-when and where they require it (see Figure 2).
Although new market entrants do not have to accommodate legacy SONET/ SDH infrastructures or OSSs, they face a different set of challenges when it comes to choosing and implementing the right metro DWDM solution. Perhaps first and foremost is speed of deployment. While the incumbent providers have the luxury of deriving revenues from SONET/SDH-based services at the same time they're deploying DWDM for niche applications, startups must put a network in place that allows them to roll out revenue-generating services right away.
In addition, if new market entrants hope to attract and retain customers, they need a metro DWDM solution that can deliver reliable service. Maximum network uptime and the ability to offer first-class service-level agreements (SLAs) are two of the characteristics that distinguish the most successful service providers in the marketplace. Network-management issues, therefore, are just as important to startups as they are to incumbent providers, albeit from a slightly different perspective at the outset.
Metro DWDM systems, regarded by some vendors and carriers as nothing more than a collection of "dumb, fat pipes," seemingly pose fewer management headaches than complex SONET infrastructures. The reality is, however, that managing any kind of transport system has always been a major challenge. Historically, the primary differentiator between network-management systems has not been Feature X or Feature Y. Rather, it has been something a lot more basic, such as does it work or not?
Therefore, to ensure they can offer reliable service right from the beginning, savvy startups want a metro DWDM solution that features a third-generation network-management system that incorporates the lessons learned from previous generations of management platforms. Although metro DWDM platforms may look like dumb, fat pipes, some startups understand they in fact are complex distributed transport systems with common "objects" running across them, which is to say they are the crucial services on which long-term success depends for both the customer and the startup.
One of the primary reasons service providers-incumbents and startups alike-are only now beginning to deploy DWDM systems in their metro networks has been initial equipment costs. To provide value-added leased services to customers, carriers, until recently, have had to provide not only the fiber, but also the associated filters and optical amplifiers that obviously represent substantial capital investments.
Now, however, some metro DWDM solutions feature a "managed fiber" option, in which each clear channel occupies one pair of fibers rather than one wavelength. Essentially, this option allows the service provider to combine space-division multiplexing (SDM) with monitoring, protection, and crossconnect features to create the first in a series of optical-channel links. Using this managed-fiber option, service providers can offer customers the value-added, DWDM -like services of a protected "virtual fiber," without having to invest in amplifiers and filters.
In fact, DWDM solutions that offer this managed-fiber option are ideally suited for service providers on both sides of an emerging metro-networking dichotomy. On the one side are service providers that opt for the more traditional networking approach, making the most of existing fiber resources by using DWDM technology to multiplex everything "into the pipe." On the other side are a small but growing number of service providers that in the interest of rapidly capturing market share are taking advantage of the lambda's ability to transport any signal format and bit rate. Using a flexible DWDM solution, such a provider can use a lambda to provide a Fast Ethernet link one day and Gigabit Ethernet the next day by merely changing the client-equipment interface and without touching the transport infrastructure (see Figure 3).
Although some providers view this "under-subscribed" networking approach as a waste of capacity, others see it as a means to upgrade their metro networks very quickly. In other words, by exploiting DWDM's interface transparencies, they can position themselves to satisfy the growing demand for capacity while at the same time simplifying their networks and reducing operational costs. Regardless of which approach a given service provider takes, the ideal metro DWDM solution can be tailored to either or both strategies.
Another issue that, until recently, slowed the deployment of metro DWDM systems has been the vulnerability of the last mile. That has been an especially critical concern in network configurations where multiple customers are connected to a single optical add/drop multiplexer. Given the frequency of fiber cuts in the access loop, this painful "Achilles heel" has prevented many service providers from deploying metro DWDM systems.
However, some metro DWDM solutions now offer last-mile protection in the form of optical customer-premises equipment (CPE). A passive optical device about the size of a pizza box, this optical CPE sits very close to the client equipment, thus requiring a very short patch between the two. Positioned so close to the client equipment, the optical CPE can supervise and protect the access loop all the way from the client equipment to the transport equipment. As a result, a service provider can readily distinguish itself from competitors by offering potential customers leased lambda services with end-to-end SLAs.
Yet another impediment to widespread deployment of metro DWDM systems has been the difficulty-and expense-of reconfiguring optical clear channels quickly and reliably. Typically, service providers could do so only by equipping each node with optical switches. Not only was that an expensive approach, but also most optical switches are not yet sufficiently robust to handle commercial applications.
To help overcome this deployment barrier, some vendors today offer a metro DWDM solution that allows service providers to use centralized crossconnecting for reconfiguring their networks. By concentrating the switching function in so-called "configurable nodes," the service provider is required to manually connect/disconnect optical inputs and outputs on only a few sites. Because these configurable nodes are centrally located and thus close to the network-management center, the service provider can reconfigure the network relatively quickly, yet not have to invest immediately in optical switches.
Although centralized crossconnecting requires more wavelengths and equipment, it is still more cost-effective than deploying optical switches in all nodes. Furthermore, it allows service providers to begin deploying optical-networking systems in metro areas with fairly stable traffic patterns. As optical-switching technology evolves, they then can upgrade their configurable nodes to achieve automatic network reconfiguration.
Business customers in all industries are taking advantage of the cost efficiencies and geographical reach of the Internet. To support their growing array of applications, they are turning to service providers for a growing array of telecom services. Basically, customers are asking their service providers for high-speed, extremely scalable, flexible, and transparent services that can be tailored to their current and future communications needs.
To satisfy these customer demands, both incumbent and startup service providers are looking for metro-networking solutions that enable them to offer cost-effective, managed transport services and top-notch SLAs. More and more service providers want to use DWDM technology to achieve the same scalability that it has provided in the long-distance market segment.
Although all service providers face the same competitive pressures in their drive to capture market share, there are enormous differences in their individual business strategies as well as in the specific service mix required by their individual customers. Incumbents and new market entrants alike realize their ultimate success in the market depends on a metro DWDM solution that offers more than scalability and transparency. They also want to work with solution vendors that can offer not just the DWDM technology itself, but also the requisite knowledge, resources, and experience to deploy a transport network.
By doing so, a service provider can obtain the ideal metro DWDM solution that has the built-in flexibility to adapt to that provider's unique strategic requirements, not just today, but also well into the future.
Shahira Raineri is vice president of marketing and Dr. Hubert Jaeger is director of solutions marketing at Optisphere Networks Inc. (Reston, VA). They can be reached via the company's Website, www.optisphere.com.