How Does Google Replace a Failed Tensor Processing Unit Rack?

Amin Vahdat says a failed tensor processing unit rack can be replaced with another rack without moving any fiber, because the light is redirected to a spare rack and the change can be done in milliseconds.

A tensor processing unit has a torus topology that connects all the units to one another directly, and a spare rack stays available at all times. Optical circuit switching leaves the bits alone in the electrical domain and, for an input port, picks which output port receives the light. The version he started with uses MEMS switches that control mirrors in three dimensions. A box might have 128 ports or 256 ports. Each input fiber maps to an output port, and rotating the mirrors sends the light to the right output port. One reason was locality between groups of racks that talk to each other a lot, such as a compute cluster and a storage cluster, using shortcuts that stay purely optical. Another reason was expanding or contracting the network and reconfiguring the spine without moving any fiber, with a controller managing the change.

Fiber stays in place rather than moving to fully free space because attenuation loss and the bandwidth would drop dramatically. Across the range of a very large building, aiming everything without fiber to connect it all together in three dimensions would be challenging and close to impossible.