The network plays a significant role in the design and development of today's large-scale distributed pervasive systems. We broadly work on network measurement, protocol design, and network-application integration. Our broad focus is to make the Internet ubiquitous and friendly for inter-connectivity among millions of devices. We list a few of our current projects here.
Video consumes the majority of the traffic on today's Internet. Users demand high Quality of Experience (QoE) from the streaming service providers; however, it is challenging to monitor and understand the state of the underlying network to stream the videos an optimal quality. We work towards making the streaming applications smarter by incorporating intelligence in the network. Such an intelligent integration between the network and the application can serve many purposes, like optimizing the video QoE, minimizing the energy consumption during video streaming, optimizing network usage, and so on. We explore specific lower-layer network properties, like the connection states, signaling information, bandwidth history, etc., to develop an online learning mechanism to optimize the video data download from the streaming server.
Next-generation 5G New Radio (NR) cellular networks operating at mmWave frequencies are targeted to support diverse use cases, such as enhanced Mobile Broadband (eMBB), massive machine-type communications (mMTC), ultra-reliable and low latency communications (URLLC), etc. Energy-Efficiency is one of the key performance indicators for NR technology. User Equipment (UE) battery life significantly impacts the Quality of Experience (QoE) of the UE. Thus the 5G NR standard is designed to have great flexibility in network operation modes to adapt to different requirements and trade-offs. 3GPP, in its 5G technical specification release, has proposed various power-saving schemes such as connected mode Discontinuous Reception (cDRX), RRC_INACTIVE state, etc. Today's ubiquitous applications need to leverage the energy-efficiency features supported by the underlying network. We primarily work in designing energy-efficient network interfaces by exploiting the application characteristics. More specifically, we target combining sensing and communication with running the applications seamlessly with fewer carbon footprints. On this front, we also contribute to developing open-source tools and simulation frameworks.