MPTCP based multi-path technology for 5G ultra-high speed and low latency services

Abstract

Various technologies have been developed for the efficient use of the multiple radio access technologies (RATs) resource at the radio access network (RAN) level or other network levels to improve user service quality in mobile communication networks. In long-term evolution (LTE), mobile carriers are commercializing RAN-level traffic aggregation technologies such as LAA-LTE, LTE-U, and LWA, which use the multi-accesses of the 3rd Generation Partnership Project (3GPP) and WiFi, and the multipath TCP (MPTCP)-based traffic aggregation technologies at the L3 network level. The standardization of 3GPP Release 16, which is scheduled to be completed by 2020, is under progress to support the traffic aggregation technology at the L3 network level through a multi-access 5G network. MPTCP is also considered as a traffic aggregation technology. However, it is difficult to apply the MPTCP employment model used in LTE to the 5G network structure as it is owing to the change to a common core architecture that accommodates multiple RATs through one common interface. Therefore, this paper proposes an optimal 5G system architecture and an MPTCP adaptation method to support the access traffic steering function based on MPTCP in a 3GPP 5G mobile communication network. We have verified the development of the MPTCP-based multi-access traffic steering (MATS) technology by implementing the proposed solution in a commercial server on a testbed based on the 5G system standard of 3GPP Release 15. Furthermore, this paper defines problems that occur when implementing the MPTCP-based MATS system and proposes relevant solutions. Based on the implementation results, it is demonstrated that the proposed MPTCP-based MATS system can perform traffic steering in the 3GPP 5G network. In the future, as low-latency services such as autonomous driving and realistic media as well as large-capacity are required, 5G multipath technology must be improved to suit low-latency services. In this paper, we propose an extended performance measurement method of ATSSS technology applicable to the proposed MPTCP-based MATS system. The proposed performance measurement method improves the delay measurement so that the MPTCP-based MATS solution can provide a low-latency service. We intend to verify the performance improvement of the proposed method through the MPTCP performance model.