Have questions? Stuck? Please check our FAQ for some common questions and answers.

You are viewing an old version of this page. View the current version.

Compare with Current View Page History

Version 1 Next »

At ON.Lab we envision future networks to consist of white boxes everywhere. This revolution was started in the data center, where people realized commodity hardware is the best way to achieve hyper scale while still guaranteeing performance and reliability. In this context, we should mention the work done by the Open Compute Project, the de facto standard for all kinds of data center equipment, including servers and packet switches. The R-CORD project has taken this one step further by driving commoditization of access network technologies, in particular fiber-based PON hardware.

It turns out that optical transport equipment is one of the last networking industries where vertical integration is the only available choice. The key component is the ROADM (Reconfigurable Optical Add/Drop Multiplexer), a type of optical switch that is usually based on WSS (Wavelength Selective Switch) technology. Today, a handful of vendors / system integrators sell chassis-based, fully integrated solutions that are typically controlled using vendor-proprietary interfaces. The TL1 protocol is most commonly used and is also around 30 years old; most vendors use TL1 and add their own proprietary messages. On top of this control plane interoperability issue, the data planes from different vendors are also incompatible, meaning a 

At ON.Lab we have built a partnership between some of the leading component and system vendors in the optical networking industry, and built the first disaggregated ROADM using open source software and commodity hardware. We disaggregated based on three ROADM functions:

  1. Transponder: the piece that takes in client side signals and generates a line side, narrow band signal. We worked with Fujitsu T100 and Ciena Waveserver.
  2. WSS or degree node: the component that performs the actual add/drop function, and controls the power level of individual wavelengths.
  3. Backplane: to connect the transponders and WSSs, we can either do static/manual connections, or use a programmable backplane. We used a 3D MEMS-based fiber switch from Calient.

Additional components that will be integrated in the near future are muxponders (consider this a more capable transponder), as well as inline fiber amplifiers, ultimately leading to a fully open line system.

In terms of control interfaces, many of the devices we use are prototypes and thus there is a lot of diversity in terms of control protocols. It is clear however that the optical networking industry is rapidly moving towards a NETCONF/YANG solution.

Our work can be considered. To achieve the vision of having both control and data plane interoperability, requires a lot of cooperation and standardization among vendors. The Open ROADM MSA, launched by AT&T, Ciena, Fujitsu, and Nokia, is a critical step in this process, that aims to achieve data plane interoperability among vendors.

  • No labels