An optical node (OOO) with no electronic conversion can be compared to a stop-free highway intersection. Traffic crossing the intersection uses ramps and roundabouts. The destination is reached without stopping. In the optical node, the data traffic passes though the node while remaining in the optical domain only, hence, maintains its autobahn-like speedy behavior.    

Today’s OEO nodes are still capable of handling and managing the data, but in face of the rapid traffic increase, such nodes will turn into throttles.    

Fortunately, the necessity for OEO conversion at networking nodes is changing. Recent advancements enable agility in the photonic layer through a combination of optical switching, gain-flattened amplifiers, ultra long haul optics, dispersion compensation, and full-spectrum tunability. These technologies make it possible for wavelengths to be networked purely ­optically up to several thousands of kilometers via intermediate multidirectional nodes, thereby bypassing the OEO transitions at intermediate nodes.

Just recently so-called Reconfigurable Optical Add-Drop Multiplexers (ROADM) and Micro-Electro Mechanical Systems (MEMS) were introduced to the market, enabling optical wavelengths routing through the core network without any OEO conversion. The concept is commonly known as optical networking or wavelength/lambda routing. The latest generations are multi-degree ROADM and offer any-port-to-any-port interconnectivity. Yet even without OEO the performance of routers is still far from being fully utilized. Switching wavelengths is ­similar to circuit switching in the electronic domain: a ­dedicated path is established first between source and destination. Using an approach related to circuit switching technology apparently has a negative impact on bandwidth efficiency, routing performance, and network utilization. Data traffic across a network is of statistically bursty nature; a behavior for which such a TDM approach (Time-Division-Multiplexing) is sub-optimal.    

Although the current generation of Multi Degree (MD)-ROADM is quite powerful in terms of bandwidth, it is nevertheless somewhat curbed in scalability due to it circuit-switched based principles and the corresponding inflexibility to respond to random traffic bursts. As a result, carriers and operators are steadily moving away from TDM towards packet switched networks where routing/switching decisions are made “on-the-fly”. Such a packet-based switching solution – never leaving the optical layer – would remove the last barriers and fully unleash the potential provided by optics. A promising candidate is Photonic Packet Switching (PPS), a family member of Photonic Integrated Circuits.