Keynote speech 3: Channel evaluation for high speed data communication

Abstract

One of the most significant things to happen in recent years is the establishment of a connected world known as “The Internet of Things” (IoT). In the IoT, the data throughput with increasing number of interconnected devices and high digital data communication for data center or personnel device is the most significant technique challenges. If the data throughput (bits/s) is not large enough, it will cause limitation of data upload and download for data communication of IoT. From the Shannon theory, the data throughput as shown in the following equation is proportional to the parameters of channel bandwidth B, number of channels M, and received signal to noise ratio S/N. How to improve the parameters of B, M, and S/N are important for the study of data throughput. In general, the various channels include wire channel and wireless channel as shown in figure 1. The common wireless channels are the antenna channel and wave propagation channel. The wire channel such as coaxial cable, twist pair cable, twinax cable, and line trace on print circuit board or semiconductor, etc.,. In order to increase the channel capacity, the high speed cable may include several channels with wider channel bandwidth and less crosstalk among channels. The common high speed data cables for data center, such as zQSFP or zQSFP-DD, include 4 or 8 channels with total data throughput 100 Gbit/s or over than 200 Gbit/s. The data throughput depends on the quality of channels. In this talk we will focus on the wire channels. There are many ways to evaluate the performance of channels. The mixed mode S-parameters for channels will be detail discussed for channel evaluation. Except for the mixed mode S-parameters of channels, the BER (bit error rate), eye pattern, and COM (channel operating margin) will also discussed during the talk.