A framework for scaling future backplanes

As line interfaces in communications chassis transition to 100 Gb/s and higher per port, many in the industry question when electrical backplanes inside these chassis will give way to optical ones. Provided that the maximum card-to-card distance over a backplane remains at one meter, two observations point to an electrical-to-optical transition near the end of this decade: commercial links transition from electrical to optical at the bandwidth-distance product of 100 Gb/s .m [1], and backplane line rates double every two process nodes and will reach 100 Gb/s in four process nodes. But optical solutions will need to overcome incumbency, or economics will call for electrical backplane solutions. To set up a framework for comparing backplane solutions, this article first defines a canonical backplane model to enumerate possible backplane implementations. Using this model, this article identifies 10 practical future backplane implementations, of which four are electrical and six optical. This article further describes how electrical solutions could be viable beyond this decade. It then outlines three necessary conditions for optical solutions to unseat the incumbent. Finally, it argues that optical solutions have an inherent advantage to scale systems with inter-chassis links, an advantage that could tip the balance in favor of optics in future backplanes.