Winter 2018 Optical Connections Magazine


Automating The Physical Layer Breaking the Fibre Barrier:

How to automate physical connectivity in fibre networks without undue impairment of optical characteristics. Robotics may be the answer to a 20-year-old problem, writes David Wang , founder & CEO, Wave2Wave Solution

C loud computing, demand for bandwidth. Thanks to the introduction of edge computing, IOT and Big Data, there is also much more fluidity as to where data resides. To meet these demands, network operators are looking for software-defined reconfigurability, fast and low-cost provision of services, improved energy efficiency, and better scalability. Software Defined Networking is a key component for a carrier to improve their network efficiency and utilisation whilst accelerating services through automation and even self- service capability for their end users. Whilst SDN is delivering great benefits and driving multi-layer optimisation across the networks, often the last piece of the automated process is sending an email to the field engineer to visit the site to make any necessary physical network changes, which marks the end of the automated process. This can introduce delays in provisioning from days to weeks or even months. The physical layer, consisting of the actual optical fibres, not only underpins the entire network, it also links access networks and provides data centre interconnects. Historically, it has been largely resistant to software definition and automation. A human operator is still required to add or remove a network connection, manually interlinking fibres on a patch panel in order to create the required network topography. The entire process is slow, from requisition to execution of service. It is also prone to error, with adjacent services at risk of disturbance during the staggering amounts of video streaming, and ever-increasing wireless communications have created an insatiable

reconfiguration of frequently jumbled patch panels. Over the years many optical switching technologies have been put forward to address the challenge without achieving any significant uptake, including 2D MEMS (Micro-Electrical-Mechanical Systems), thermo-optic, and liquid crystal. All three of these technologies have scalability constraints, with port count restricted to 32x32 for reasonable switch insertion loss. Large port count optical switches do exist, but have their own limitations and drawbacks. Using multiple 3D MEMS optical switches, larger port count switches can be built. However, the optical losses of different connections in a 3D MEMS optical switch are intrinsically different, resulting in a relatively large insertion loss spread (up to 3-4dB) and the total optical loss quickly becoming unbearable for many applications.

with lens collimators and arranged into 2D arrays facing each other. Every fibre collimator is mounted to a 2D Piezo actuator which controls the collimator beam pointing in free space. To make a connection between an input fibre port and an output fibre port, the 2D Piezo actuator steers the input fibre collimator to point its beam to the destination collimator. With proper fibre collimator design, low insertion loss of ~0.4dB can be achieved. However, the path lengths of different fibre to fibre connections vary. The collimators can only be optimised for one given pathlength, thus an insertion loss variation exists across different connections. Non-uniformity of the fibre collimators also widens the loss spread. As a result, commercial switches using direct beam steering are specified with a 1.2dB typical loss and a 2.2dB maximum loss. Switching can be relatively fast, approximately 20ms. However, similar to 3D MEMS optical switches, there is no locking mechanism for the established connections and, therefore, it is not suitable for mission-critical applications. Fibre coupling using direct beam steering in theory can support both


One optical switching technology with commercial success is direct collimator steering. As shown in Figure 1, both the input and output fibres are terminated

Figure 1: Conceptual view of a collimator steering based optical switch


| ISSUE 15 | Q4 2018

Made with FlippingBook - Online catalogs