These are exciting and challenging times for network operators. The growing impact of broadband everywhere portends huge changes in the network. For starters, traffic volume continues to increase as more people make more use of high-bandwidth applications. The nature of that traffic is also evolving with growing transport of jitter- and delay-sensitive triple-play services, including voice and video. To top it all off, the network must increasingly accommodate the kinds of bursty traffic flows that “on demand” services, such as video-on-demand and videoconferencing, can create. These types of services require the highest quality-of-service levels to satisfy bandwidth requirements and response times.
Simply “adding more bandwidth” won’t meet these challenges. Bandwidth prices are extremely sensitive to competition and have been falling every year since 2001. With operating expenditures (opex) making up 70% to 80% of cost, network operators must dramatically increase bit rates and aggressively decrease opex to maintain a positive return on investment.
Network operators must examine every angle to ensure the ongoing viability of their business. All resources, both human and network, must be optimized to ensure the network operates at peak efficiency and will support growth and change. At the optical transport level, reconfigurable optical add/drop multiplexers (ROADMs) can improve capacity and maximize efficiency when compared with fixed OADMs.
Conventional fixed OADMs drop wavelengths by using preselected fixed-wavelength filters; they optically pass through all other express wavelengths.
Setting up or reconfiguring an OADM network can require weeks of work, as specialists travel from one site to another carrying out precise, manual procedures, including configuring each node, analyzing and balancing optical power, and tuning lasers. In many cases, when the network needs to be adjusted, this manual process must be repeated. Obviously, this is expensive and hampers the network operator’s ability to roll out services quickly. It can also cause significant service disruption for existing services.
The lack of flexibility in a fixed OADM system also contributes to bandwidth inefficiencies. Operators must make certain assumptions when setting up the system that may not be correct or that become obsolete as traffic flows evolve in an unforeseen direction. Bandwidth also ends up stranded or unavailable due to the fixed OADM’s inability to manage bandwidth at the granularity of a single wavelength, as it typically adds and drops predetermined groups of wavelengths.
For these reasons, network operators around the globe have begun to recognize the value of ROADMs. Recent reports indicate that 65% of all network operators plan to integrate ROADM technology into their metro networks. The ROADM segment is growing by a compound annual growth rate (CAGR) of 65%.1
A ROADM remotely controls the adding, dropping, and passing through of wavelengths without converting the optical signal to the electrical domain. ROADMs support any-to-any port connectivity without reengineering, and they offer a level of flexibility at the optical layer that is similar to what SONET/SDH ADMs provide at the subwavelength level.
ROADMs lower the total cost of ownership by reducing opex and increasing the network operator’s ability to get services up and running quickly. They are ideal systems for today’s network because they enable:
• Simple, fast service activation through a fully automated optical layer.
• Flexible network configurations that can be adapted as requirements change.
• Improved bandwidth efficiency by eliminating stranded or underutilized bandwidth.
• Integrated SONET/SDH and WDM layers that simplify the network.
Before adopting a ROADM architecture, network operators should consider the specific capabilities of the different ROADM types. These types are usually identified by the technology they use to perform the wavelength filtering and switching functions: discrete, wavelength blockers, wavelength blockers with integrated planar lightwave circuits (IPLCs), and wavelength-selective switches.
Early ROADM technology was based on discrete optical-mechanical switches, filters, and variable optical attenuators (VOAs). While simple to implement, since most of the technology was commercially available, this approach used many optical components. The result was very high insertion loss, very high cost, and a large footprint-which prevented widespread acceptance of this approach.
A wavelength blocker architecture splits the incoming DWDM signal into a drop and through path. An integrated DWDM demultiplexer, VOA, and multiplexer form the core of the wavelength blocker. Typically, blockers are implemented using microelectromechanical systems (MEMS) or liquid-crystal technologies.
While this architecture reduces the number of discrete components, it forces the network operator to pay for all wavelengths at each node from day one. Furthermore, this design only manages the through wavelengths, not the add or drop wavelengths. Most implementations of this architecture use inflexible fixed filters for the add and drop wavelengths, so this design is really only a “semi-reconfigurable” OADM.
A variant of the wavelength blocker architecture uses IPLCs. This blocker and IPLC architecture integrates the add multiplexer with the wavelength blocker. This eliminates the extra add multiplexer, but at the expense of requiring additional optical switches. Furthermore, these designs permanently lock the add wavelengths to a fixed wavelength design, which makes fully tunable lasers only good for sparing (since they are not being used for dynamic wavelength management).
IPLC-based designs do not inherently support growing to greater than two-degree node applications (east and west transmission), an important desired feature for interconnecting rings and subtending spurs.
The best ROADM architecture available is a third-generation ROADM system based on a multiport, wavelength-selective switch (WSS) architecture (see Figure 1). This architecture lives up to the “R” in ROADM, enabling the network operator to respond to changing network requirements on the fly. Wavelength switches can direct one or more wavelengths from an incoming DWDM signal to one or more output ports, usually with individual VOA-like power control for each wavelength via dynamic gain equalization (DGE) filters.
The WSS-based ROADM enables a fully automated optical layer, with single-wavelength granularity, to deliver both opex and capex savings. The WSS ROADM performs four functions in one unit:
• Fully flexible optical add/drop with colorless ports.
• Dynamic management of received power.
• Dynamic power equalization.
• Continuous optical monitoring of all channels.
In addition to the WSS ROADM module, the ideal third-generation ROADM-based system will include:
• Full C-band tunable lasers to support dynamic wavelength management and reduce the cost of sparing.
• Variable gain, transient controlled amplifiers to ensure the robustness of the network.
• A fully automated GMPLS control plane to ensure painless operations and provisioning.
• Colorless ports with preinstalled jumpers to eliminate hours of manual setup time and the risk of errors connecting patch cords.
• Multiservice interfaces to aggregate lower-speed services onto a wavelength (e.g., multiplex nine Gigabit Ethernet signals onto a 10-Gbit/sec wavelength).
• Intelligent optical and electrical switching for the most efficient bandwidth utilization.
• Planning tools that enable one-time network engineering, with the ability to readily accommodate future growth without reengineering established elements.
• A modular architecture that enables a lower initial setup cost and supports non-service-affecting growth.
ROADMs have been deployed effectively in optical networks since 2003. The following examples provide insight into how different market segments have taken advantage of ROADM features to improve their bottom lines.
A large university’s installed optical network couldn’t handle its growing voice and data traffic, but the university’s network engineers wanted to avoid laying new fiber. They needed a solution that would allow them to get more out of their existing infrastructure. By implementing a ROADM, the university dramatically increased the network’s capacity and added the ability to change and expand services with minimal effort and cost.
With a ROADM-based network, the university enjoys these benefits:
• Flexible infrastructure-the new multiservice network carries both SONET and Ethernet traffic over a common infrastructure (see Figure 2).
• Efficient bandwidth utilization-the use of DWDM has significantly improved network capacity, enabling the university to support the growing demands of its users without laying additional fiber.
• Ability to add services quickly-a WSS ROADM makes it very easy to adapt services on the fly, a necessity in a research environment where needs change dramatically as some projects gear up and others reach completion.
In another example, an infrastructure service provider needed to expand its metro transport network to meet growing demand from its customers for more and faster bandwidth. In addition to modest growth in SONET traffic, the service provider was experiencing very high growth in Gigabit Ethernet traffic. The provider needed an approach that would enable it to control operating costs while meeting this increased demand (see Figure 3).
The implemented ROADM network provides several benefits. It improves network efficiency via the ability to aggregate multiple Gigabit Ethernet streams onto a 10-Gbit/sec wavelength. The ROADM’s use of pluggable optics and multiservice/multirate line cards provides per-port service flexibility with minimal line card investment. The network also enables the service provider to realize operational savings through automated remote provisioning of wavelengths that are as easily provisioned as SONET circuits. Finally, it simplifies network operation because the service provider can add capacity, wavelengths, and even additional nodes to the network without having to reengineer the entire optical ring.
A full-featured, third-generation ROADM may not appear to be the most cost-effective approach if all one looks at is the initial equipment cost. However, with the huge majority of the cost of a network coming from the ongoing operation of that network, it makes sense to base decisions on the total cost of ownership. When capex and opex are weighed together, ROADMs easily stand out as the most cost-effective option. From installation to day-to-day operation, growth, and evolution, third-generation ROADMs help ensure the ongoing profitability of the network operator’s business in dynamic times.
Steve D. Robinson is vice president of product management at Meriton Networks (Ottawa, ON, Canada; www.meriton.com). He has more than 20 years of business and technical experience in the telecommunications industry with companies such as Agere Systems, Lucent, and AT&T Bell Laboratories. Robinson received both his bachelor’s and master’s degrees in electrical engineering from Drexel University and his master’s in business administration from Lehigh University.
1. Determined from data sources that include Infonetics Research, Heavy Reading, and Meriton Networks.