The Optical Backplane is Finally Here


Will This Change Everything


The mythological optical backplane has been in the works for 20 years. By the late 1990’s it was being readied for the time when copper ran out of speed which was always sometime in the near future. This author was part of a DARPA-sponsored optical backplane project called POINT (Polymer Optical Interconnect) launched in 1996. With several partner companies, we demonstrated a fully functional backplane over 52 cm of flexible polymer waveguides with disconnects included. It worked very well.

Subsequent to this, countless demonstration projects have been and continue to be developed and reported. Europe has been a major contributor with many pre-competitive consortia projects for optical backplanes.

In the 1990’s and 2000’s, fiber to the home was always just three years away. It was always a moving target. But FTTx finally achieved mass deployment will many millions of connections deployed, changing broadband deployment in a big way, and enabling the emergence of the digital economy. Is the optical backplane finally ready to change things inside high-performance systems?

What has taken so long?

Why is it that a set of optical technologies that were viable by the year 2000 are still not mainstream? The answer is CMOS. Advances in CMOS have changed everything in data communications. Signal conveyance over copper backplanes at 10Gbps was considered challenging years ago. Today, 25G backplanes are being deployed using advanced signaling techniques, including FEC, though many would still label 25G as challenging. Serdes are being announced for 50G and standards will soon be ready. At OFC, the OIF held a workshop on 100G electrical signaling.

Another factor working against optical backplanes is simple economics. For years, system architects have suggested they will use optics as a backplane when the cost meets that of copper, but that will not happen in the near future. The components of an optical link are not as low cost as a copper trace and a couple of CDRs (clock and data recovery chips). So the choice between copper and optical backplanes comes down to need. At 100G, even if a copper backplane can be built, it will be power hungry. Will that be the point where we need an optical backplane?

To be fair, both Cisco and Juniper Networks have been shipping core routers based on intra-system optical interconnects for many years. One could fairly call these optical backplanes across multiple chassis. Fabric chassis are connected to line card chassis with massive numbers of optical interconnects. But each chassis still has a conventional copper backplane.

Each year, LightCounting assesses the outlook for optical backplanes in our report on embedded optical modules ( The necessary technologies are in place, and wide-scale adoption will occur when speed, interconnect density and power limitations align to exceed the limits of copper. Planning for system architectures that will live into the future can also swing the balance to optics.

For two companies, Oracle and Ericsson, these stars have aligned to put systems into production based on optical backplanes using embedded optical modules or EOMs (a.k.a. on-board optics, mid-board optical modules, board optical assemblies, etc.).

LightCounting clients can access the full version of the research note at:





LightCounting is a leading optical communications market research company, offering semiannual market updates, forecasts, and state-of-the-industry reports based on its analysis of primary research with dozens of leading module, component, and system vendors as well as service providers and other users. LightCounting is the optical communications market’s first choice source for the accurate, detailed, and relevant information necessary for doing business in today’s highly competitive environment. For more information, visit: or follow us on Twitter at @LightCounting.