Welcome to Lightning, the next-generation photonic computing system invented here at MIT!
Lightning is a reconfigurable photonic-electronic smartNIC that serves real-time deep neural network inference requests at 100 Gbps. Lightning uses a novel datapath that feeds traffic from the NIC into the photonic domain without creating digital packet processing and data movement bottlenecks. Lightning achieves this by employing a reconfigurable count-action abstraction, which keeps track of the computation Directed Acyclic Graph (DAG) of each inference packet. Our count-action abstraction decouples the compute control plane from the data plane by counting the number of operations in each task of the DAG and triggers the execution of the next tasks as soon as the previous task is finished, without interrupting the flow of the data.
 Delocalized Photonic Deep Learning on the Internet's Edge
Alexander Sludds, Saumil Bandyopadhyay, Zaijun Chen, Zhizhen Zhong, Jared Cochrane, Liane Bernstein, Darius Bunandar, P. Ben Dixon, Scott Hamilton, Matthew Streshinsky, Ari Novack, Tom Baehr-Jones, Michael Hochberg, Manya Ghobadi, Ryan Hamerly, Dirk Englund