Dissertation Defense

Design techniques for high performance wireline communication and security systems

Shiming Song


The amount of data traffic grows exponentially on the internet, towards thousands
of exabytes by 2020, calling for innovative solutions. Within-server communication dominates today's network data transfer, outweighing between-server and server-to-user data transfer by an order of magnitude. Solutions for within-server communication tend to be very wideband, i.e. on the order of tens of gigahertz, and equalizers are widely deployed to provide extended bandwidth at reasonable cost. However, equalizers typically cost the available signal-to-noise ratio (SNR) at the receiver side. This thesis presents two equalization solutions to address the optimal equalizer implementations. One solution is a low-power high speed maximum likelihood sequence detection (MLSD) that achieves record energy efficiency of sub-10 pico-Joule per bit. The other solution is a phase-shaping equalizer that suppresses inter-symbol interference at almost zero cost of SNR. The growing amount of data communication also poses security challenges. The emergence of quantum computers demands advanced cryptographic solutions to enhance data protection. With many attractive features including post-quantum security, lattice encryption offers great promises. This thesis presents a fast and efficient programmable lattice encryption processor that surpasses the energy efficiency of state-of-the-art solutions by 4 orders of magnitude, opening the door to many future embedded applications.

Sponsored by

Professor Zhengya Zhang