Cloud Native Solutions for Orchestration in LTE Networks

P C, Amogh and Tamma, Bheemarjuna Reddy (2018) Cloud Native Solutions for Orchestration in LTE Networks. Masters thesis, Indian Institute of Technology, Hyderabad.

Text
Thesis_Mtech_CS_4035.pdf - Accepted Version
Restricted to Repository staff only until July 2021.
Download (2MB) | Request a copy

Abstract

Due to rapid growth in the use of handheld mobile devices and as a vital carrier of IoT traffic, mobile networks must undergo necessary infrastructure-wide revisions to alleviate the traffic explosion. In addition to data traffic, there has been a significant rise in the control signalling overhead due to more handover under small cell scenario and IoT. Adoption of technologies like cloud computing, Software Defined Networking (SDN) and Network Functions Virtualization (NFV) is impressively successful in mitigating the existing challenges and driving the path towards 5G evolution. However, issues pertaining to scalability, ease of use, service resiliency, and high availability need considerable study for successful roll out of production grade 5G solutions in cloud. Taking advantage of the advances in network virtualization, using NFV, SDN and the Cloud computing, the 5G system is making network slicing a reality. 3GPP as well as 5G Infrastructure Partnership by European Commission has come up with several 5G architectures that support network slicing. They classify slices into three categories (i) massive Machine-Type Communication (mMTC) for IoT services, (ii) Ultra-Reliable Low-Latency Communication (URLLC) for low latency communication, and (iii) extreme Mobile Broadband (xMBB), which requires high data rates. We propose a scalable Cloud Native Solution for Mobility Management Entity (CNS-MME) of 5G mobile core in a production data center based on microservice architecture. The proposed architecture is highly available and supports auto-scaling to dynamically scale-up and scale-down required microservices for load balancing. We discuss the implementation of a Network Slicing Engine for autoscaling of data planes in LTE core network. We also propose a RAN Slicing framework which can schedule downlink and uplink transmissions of network slices. We focused on reducing latency of URLLC slice by applying scheduling policies. Our scheme ofiers strict requirements, especially in terms of latency and reliability as shown in our simulation results.

[error in script]

IITH Creators: