An optimization method for accelerating TCP convergence in metaverse scenarios

In recent years, the rapid development of virtual reality (VR) technology has brought the concept of the metaverse from science fiction to reality. VR devices have made significant progress in the consumer market, with new hardware and technologies constantly emerging. However, the initialization time for VR application software is quite long, severely impacting user experience. We closely examined the Transmission Control Protocol (TCP) connection behavior during software startup and found that current congestion control algorithms take a long time to converge to the maximum available bandwidth in highbandwidth networks after the slow start phase. This is because the internal parameters of widely used congestion control algorithms, including CUBIC, are fixed within the kernel, leading to prolonged convergence times in high-bandwidth environments. In this paper, we propose FWMB, a dynamic parameter adjustment method based on CUBIC that optimizes parameter settings using the current network’s maximum available bandwidth. We implemented this algorithm in a Linux environment, and comprehensive experiments show that compared to the default CUBIC algorithm, FWMB reduces the convergence time to maximum bandwidth by $11.2 \%$ to $19.7 \%$.