Considering reconfigurable optical systems
Carriers can no longer afford to build networks based on contentious bandwidth projections alone
BY MOULI RAMANI
Polychromix
Wilmington, MA
http://www.polychromix.com
The optical telecommunications market has changed dramatically over the past few years. By selecting devices that enhance the efficiency of networks, you can simultaneously lower operating and capital expenditures.
The introduction of dense wavelength division multiplexing (DWDM) has enabled carriers to dramatically increase the data carrying capacity of existing fiber at a lower additional cost by separating light signals into tightly spaced wavelengths, each carrying a separate data signal.
Many optical networks cannot be automatically reconfigured, making it extremely difficult to provision additional wavelengths or route signals to a different location. Doing so requires de-multiplexing of all individual wavelengths followed by the replacement or redistribution of dozens of static components (for example, gain-flattening filters or GFFs) that are customized for a specific network state.
This process increases cost in time and money and requires manpower to alter the network configuration, while network management complexity grows significantly as new wavelengths are added. More than any other factor, these complexities are responsible for the long and complicated process of network upgrades.
Operational considerations
Dynamic reconfigurable devices can compensate for many of the shortcomings of fixed components by enabling higher quality, balanced light transmission. Such improvements enable the increase of distance between optical amplifiers, reduce the need for costly electrical regenerators and eliminate the repeated cost and requirement for customized gain filters, demultiplexers and variable optical attenuators (VOAs).
Currently, provisioning of any new wavelength requires expensive and labor-intensive truck rolls for removal and replacement of the aforementioned Gain Flattening Filters and associated equipment as well as system reconfiguration. Additionally, connecting new circuits often involves manual patch-panel reconfiguration and extensive network downtime that often result in long periods of delayed services and lower customer satisfaction.
To address this challenge, service providers are evaluating end-to-end management systems that support remote provisioning, activation, and fault correction. Such capabilities enable carriers to redirect capacity and rebalance their networks using keystrokes from a NOC rather than requiring technician truck rolls to network nodes and amplifier sites.
Dynamic optical-power balancing and routing solutions bring differing benefits to various optical telecom segments including metro, long haul (LH) and ultra long haul (ULH). While a key benefit in LH and LH networks may be correction for amplifier gain tilt and enabling light to travel farther at a cheaper cost, in-metro networks the need may be driven by frequent adding and dropping wavelengths and the non-uniformity of these wavelengths' power levels.
System enhancements
Dynamic reconfigurable networks enhance a system by providing the ability to develop new wavelength-based services and more efficiently manage bandwidth. Such systems eliminate the constraints on current network infrastructures.
Many enterprises are required to lease a wavelength from a carrier despite using the capacity for a fraction of the time leased, as network bandwidth is over-allocated to accommodate services and applications. By enabling dynamic provisioning, carriers could utilize their bandwidth in a much more efficient manner, shifting capacity on-demand and thus extracting more bandwidth from current networks.
Another bandwidth allocation constraint is the time to market for provisioning new network capacity, a problem that becomes even more acute in cases of irregular traffic patterns. Dynamic networks eliminate the need to start months in advance in order to provide additional bandwidth.
Dynamic spectral equalizers (DSEs) integrate subsystems incorporating optics and electronics to dynamically adjust the optical power spectrum of a DWDM signal and transform it into a desired target spectral shape. Such functionality is important in any optically amplified system as optical amplifier gain uniformity depends on optical power and its distribution across the amplification band.
Current networks use static GFFs to balance tilt created by optical amplifiers. One GFF shortcoming is that they are designed to flatten the gain of a fully loaded amplifier. In most cases however, only a few channels are lit in a system at deployment time, leaving unlit spectrum for later use either by newly originated wavelengths or by wavelengths routed from other locations to the amplifier site.
Dynamic spectral equalization (DSE) becomes important is such cases of change in the network since varying optical power will affect the amplifier's gain uniformity. Static filters cannot compensate for these dynamic spectral changes. Next-generation dynamically reconfigurable networks will require DSE to replace GFFs, providing a scalable solution that is agnostic to wavelength distribution within the network.
The devices are designed in many cases to work in a “closed-loop” fashion, meaning that a control signal is constantly sent back from an optical power monitor to the DSE for correction. Such open-loop capable devices with 50-GHz spaced control points across the entire C band are readily available in the market, such as the Polychromix Dynamic Channel Orchestrator 10050A.
Dynamic channel equalizer
Similarly to the DSE, the dynamic channel equalizer (DCE) controls power on a per channel basis, however, rather than generating an averaged, smooth spectral slope, the DCE maintains a wide passband for each wavelength. This approach enables some DCE technologies to provide full attenuation of any combination of channels to a fully blocked state.
Attenuation of over 45 dB enables the use of DCEs as wavelength blockers, which are useful in creating broadcast and select ROADM architectures in addition to the power balancing functionality described above. There is a growing trend in having the capability to fully control the power level of each individual channel in a WDM system rather than averaging its power level with neighboring channels.
Such spectral isolation is now available for up to 100 wavelengths in subsystems such as the Polychromix Dynamic Channel Orchestrator 10050AB (see Fig. 1 ). ROADM configurations of this nature are much more flexible than static ROADMs that require manual adjustments for any provisioning event, and are much more cost effective than combining de-multiplexers, large switch arrays and multiplexers.
Spectral isolation devices such as the Dynamic Channel Orchestrator provide spectral isolation for up to 100 wavelengths.
Wavelength selective switches
A next product type that is still in its development phase is the wavelength selective switch (WSS). These devices can switch any wavelength in a WDM transmission into any of several output ports. Another common name for such products is the wavelength-selective 1 x n switch as wavelengths from one input port can be switched to n different output ports.
Such devices provide a building block that enables the creation of ring interconnects as well as the ability to cost effectively transform to mesh architectures. The market has yet to converge on the specific requirements from such devices, although it is clear that they will play a substantial role in next generation network architectures side by side with wavelength blockers for wavelength routing applications.
Many technologies are racing to deliver the functionality of these devices. Although no technology has emerged as the clear winner, many systems houses and carriers believe that micro electro-mechanical systems (MEMS) will be a major part of the integration effort to create low-cost low-loss OADMs and tunable devices.
Functionality and performance
The first thing to consider is optical performance, as a high channel-count with a 50-GHz ITU grid spacing enables carriers to grow their channel count as needed without having to upgrade system equipment. A small spectral gap between wavelengths enables banding of tighter or looser spaced wavelengths. (Technologies such as diffraction-based MEMS are capable of such performance.)
Low insertion loss is important, and a wide dynamic range of attenuation is required when blocking a channel (attenuation of greater than 40 dB is commonly considered adequate.) Tight optical specifications also minimize PDL, CD and other optical degradation effects.
A small footprint (single line-card) increases system density, and low power consumption enables a carrier to save electricity and reduce the heat-related problems. System houses and carriers require sub-system manufacturers to incorporate functionality of multiple network elements into one single device.
Advertisement
