Bulletin No. 2, 2024

stations. “When two ‘parcels’ are sent to the same intersection, they are stored and categorised before being forwarded to the next node,” as Professor Yeung describes it. In the context of wireless networks, the more nodes that the parcels pass through, the more serious the data loss becomes. “Increasing the transmission power of wireless signals to extend the transmission range could reduce the number of relay nodes. But it will interfere with the signals of other types of wireless communication,” he adds. Network researchers have been looking for solutions to reduce data loss without affecting transmission power. He left AT&T Bell Laboratories, a premier US research laboratory, in 1991 to join CUHK’s Department of Information Engineering, where he delved into network research. Researchers previously believed the best way to send a message through the network was to encode it into data packets for high-speed delivery, without altering the packets inside the network. But he proved that mixing data packets inside the network could transmit more information. He coined the term “network coding” to describe it. “I spent seven to eight years working with collaborators at CUHK and overseas to develop the theory.” His foresight and perseverance gave birth to a paradigm shift, enabling more information to be transmitted through networks at a faster rate, so that, for example, data can be downloaded faster from the internet and video streamed with less delay. The butterfly effect that trumps traditional routing With the aim of maximising the theoretical throughput of network communication, Professor Yeung and his colleague Professor Robert Li Shuo- yen came up with a novel example called the “butterfly network”, in which they showed that with network coding, information can be transmitted faster than with the traditional “store-and-forward” paradigm. “The data are sent as coded packets in the butterfly network. The data packets received at the intermediate network nodes are encoded into new packets that are sent to the next node, where the original message is reconstructed at the destination nodes,” he says. The technology later evolved into BATS code, which is a joint effort between Professor Yeung and his former PhD student Yang Shenghao, now a professor at the School of Science and Engineering at CUHK-Shenzhen. BATS code simplifies the encoding and decoding processes to increase network throughput, reduce delay and make the network more robust. Their project, “Network Coding for Next Generation Networks”, was recognised by the government’s RAISe+ Scheme this year. | Professor Raymond Yeung Chinese University Bulletin 28