Question

Regarding TCP in 5G/mmWave networks, the 5G that may introduce links with extremely variable capacity and retransmissions of the packets will face problems. What could be possible solutions for TCP in 5G?

Answer 1

The vast available spectrum in the millimeter- wave (mmWave) bands offers the possibility of multi-gigabit-per-second data rates for fifth generation cellular networks. However, mmWave capacity can be highly intermittent due to the vulnerability of mmWave signals to blockages and delays in directional searching. Such highly variable links present unique challenges for adaptive control mechanisms in transport layer protocols and end-to-end applications.

Some recent studies have highlighted that the extreme variability of the signal quality over mmWave links yields either degraded TCP goodput and very low utilization of the resources at mmWave frequencies, or, in the presence of link-layer retransmissions, high goodput at the price of high latency. While the TCP goodput degradation caused by buffer overflow in undersized buffers is difficult to mitigate, the problem of bufferbloating (i.e., large buffer occupancy leading to delays) can be approached by appropriately designing cross-layer algorithms.

In order to assess how TCP will perform in mmWave cellular networks, we consider two of the most challenging scenarios among those specified by 3GPP, that is, a high-speed train and a dense urban scenario. Even if blockage events are still possible, in a scenario with a dense deployment (according to 3GPP guidelines), it is possible to provide uninterrupted connectivity to the final user.

Despite the advancement ofbeamforming technology, the following problems remain inthe mmWave band.

•Blockage problem: The phenomena that the signal can-not pass through the obstacle owing to the directivity andthe receiving SNR value is severed.

•Beam misalignment: The receiver has a poor SNRvalue owing to nonmatching of transmitting/receivingbeam pairs.

In particular, the use of the transmission control proto-col (TCP) is essential owing to the characteristics of disaster communication, where end-to-end communication reliabilityis very important. However, in 5G networks, especially in mmWave band, TCP is very vulnerable owing to beam mis-alignment and blockage issues. The conventional loss-based TCP, such as Haet al, is not suitable for the mmWave capacity owing to its wide bandwidth, while congestion-based TCP, such asBBR, is more suitable.

The proposed TCP performs learning based on mobility, location, and reception signal-to-noise ratio (SNR) value of the terminal, and the learning agent (UE)predicts the time duration for which the transmitting signal is disconnected (blockage duration). If the blockage duration is less than the TCP re-transmission timeout (RTO) value, the congestion window (cwnd) value is fixed and buffering is performed for the corresponding time to utilize the mmWave capacity. By contrast, the congestion control (CC) algorithm is performed by determining that the actual congestion has occurred.

You can read https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=8859212 for a better understanding.