The electronics industry in the last four decades, approached the circuit and system design from the perspective of building the system with the lowest power, lowest area, and highest speed irrespective of the application. However, with Internet-of-Things and Big data picking up, the next generation of computing and communication system is moving towards having - a cloud where all the computing and storage are done, a hand-held or dummy terminal that communicates with the cloud at very high data rates to get all the required information, and the usual base-station which forms the backbone of the communication network.
CMOS is finding it difficult to meet all the requirements of such systems and may no longer be the optimal and unique solution for building all parts of such systems. Time may be right to approach the overall system design using the “Divide-and-Conquer” methodology – build the storage and compute resources in the cloud and base-stations using the best possible technology for the maximum speed, and highest storage density, but not necessarily the lowest power. This technology giving us the speed and storage density benefit needed for cloud and base-stations could be a hybrid solution. The remaining parts of the system, where power is a critical requirement, can be built using CMOS technology or by using other low-power hybrid solutions.
This talk will focus on my past research experiences and future ideas of various hybrid analog and mixed signal solutions that not only use different technology nodes but also use signal