PaperlayerView exact revision on arXiv

Source-linked AI summary

Fog Radio Access Networks: Mobility Management, Interference Mitigation and Resource Optimization

Haijun Zhang, Yu Qiu, Xiaoli Chu, Keping Long, Victor C. M. Leung

arXiv:1707.06892v1cs.ITmath.IT

TL;DR

HCRAN leaves its cost structure, BBU-supported resource optimization, and cloud-center traffic burden in need of further study. The paper presents FRAN with edge caching, computing, signal processing, and resource management, and reports reduced handover signaling overhead alongside interference mitigation and increased net utility.

  • Problem

    HCRAN's cost structure, BBU-supported resource optimization, and traffic burden on the cloud center require further study.

    Source: dozens of network standards have… · p. 2

  • Method

    The paper presents an FRAN architecture with edge caching, computing, cooperative radio signal processing, and resource management, alongside handover and game-based interference/resource mechanisms.

    Source: §V · p. 14, have proposed mobility management… · p. 15

  • Results

    The handover scheme significantly decreases handover signaling overhead, and the game mitigates cross-tier interference and increases total net utility.

    Source: have proposed mobility management… · p. 15

  • Takeaways & Limitations

    FRAN places caching, computing, cooperative radio processing, and resource management at the network edge.

    Source: §V · p. 14

Abstract

from arXiv · show

In order to make Internet connections ubiquitous and autonomous in our daily lives, maximizing the utilization of radio resources and social information is one of the major research topics in future mobile communication technologies. Fog radio access network (FRAN) is regarded as a promising paradigm for the fifth generation (5G) of mobile networks. FRAN integrates fog computing with RAN and makes full use of the edge of networks. FRAN would be different in networking, computing, storage and control as compared with conventional radio access networks (RAN) and the emerging cloud RAN. In this article, we provide a description of the FRAN architecture, and discuss how the distinctive characteristics of FRAN make it possible to efficiently alleviate the burden on the fronthaul, backhaul and backbone networks, as well as reduce content delivery latencies. We will focus on the mobility management, interference mitigation, and resource optimization in FRAN. Our simulation results show that the proposed FRAN architecture and the associated mobility and resource management mechanisms can reduce the signaling cost and increase the net utility for the RAN.

I. INTRODUCTION

Centralized RAN architectures face mobility, traffic, and resource-management challenges, while FRAN distributes caching, computing, signal processing, and resource management at the network edge. The article presents FRAN and associated mechanisms for handover, interference, and resource management.

dozens of network standards have emerged, including the Third Generation Partnership Project

  • Centralized CRAN and HCRAN architectures make it difficult to use social and local information in RAN management and optimization.HCRAN also raises questions about resource optimization support and traffic burden on the cloud center and fronthaul/backhaul links.Source: dozens of network standards have… · p. 2
  • FRAN uses distributed edge computing and storage rather than relying solely on a centralized BBU pool.The article describes local signal processing, cooperative resource management, and distributed caching at the network edge.Source: dozens of network standards have… · p. 3
  • The article proposes FRAN handover management and interference/resource mechanisms, including pricing and game-based subchannel and power allocation.The allocation problem is modeled as a non-cooperative game.Source: dozens of network standards have… · p. 4
  • The proposed interference and resource management mechanisms reduce interference in FRAN and enhance the network's net utility.The introduction reports these outcomes without giving quantitative values.Source: dozens of network standards have… · p. 4

II. SYSTEM ARCHITECTURE

FRAN connects macro and small remote radio heads and fog-computing access points to a BBU pool, while assigning the edge distinct processing and caching roles. Local caching lets users obtain requested content without connecting to the BBU pool each time.

II. SYSTEM ARCHITECTURE

  • F-APs integrate fronthaul radio functions with physical-layer and upper-layer processing, enabling local collaborative signal processing and flexible cache management.SRRH and F-AP caches serve different content: F-APs store highly locally popular or relevant content, unlike SRRHs.Source: §II · p. 4
  • F-APs can cache popular content to serve nearby users locally, so users need not connect to the BBU pool for every data or content request.The text relates this role to neighboring users' similar social-application traffic and shared content requests.Source: §II · p. 4
  • MRRHs serve high-mobility F-UEs, while SRRHs and F-APs provide high capacity to a smaller number of low-mobility F-UEs.MRRHs connect to the BBU pool over backhaul; SRRHs and F-APs use fronthaul links.Source: §II · pp. 4–5

caching capabilities to provide local caching services to their associated F-UE. Collaborative

FRAN assigns local signal processing and distributed resource management to edge devices, including SRRHs, F-APs, and F-UEs. The article describes multiple transmission modes and user-centric adaptive selection among them.

caching capabilities to provide local caching services to their associated F-UE. Collaborative

  • SRRHs, F-APs, and F-UEs can perform signal processing locally, while F-APs and F-UEs manage resources in a distributed manner.This distribution can relieve fronthaul communications burden and BBU-pool processing and control burden.Source: caching capabilities to provide local… · p. 6
  • Edge devices receive functions including distributed storage, control, configuration, measurement, and resource management.The article presents these as new functionalities assigned to network-edge devices.Source: caching capabilities to provide local… · p. 6
  • FRAN offers D2D and relay, local distributed coordination, global centralized, and MRRH transmission modes.F-UEs select a mode using adaptive techniques that account for movement speed, distance, location, QoS, computing capability, and caching capability.Source: caching capabilities to provide local… · p. 6

A. Handover Management

The article identifies frequent or unnecessary handovers as a challenge in heterogeneous networks and proposes shifting some handover control and decisions to F-UEs. It also distinguishes service by mobility: MRRHs for high-speed users and SRRHs or F-APs for low-mobility users.

A. Handover Management

  • High-speed F-UEs should use MRRHs, whereas low-mobility F-UEs should use SRRHs or F-APs.The text characterizes MRRHs as offering large coverage and reliable connections, and SRRHs or F-APs as high-capacity options for fewer users.Source: A. Handover Management · p. 6
  • Small-cell coverage and severe co-channel interference make unnecessary handovers and radio-link-failure-related handovers more likely in heterogeneous networks.Frequent handovers also burden the fronthaul, backhaul, and core network.Source: A. Handover Management · p. 6
  • MRRHs, SRRHs, and F-APs can shift some handover control and decision-making to F-UEs to avoid frequent handovers and alleviate handover control overhead.The passage frames this as a way to address handover management in FRAN.Source: A. Handover Management · p. 6

B. Handover Procedures in FRAN

Handover management in FRANs differs from that in heterogeneous networks.

or in CRANs, the handover procedure should be revised for both F-AP/SRRH-to-F-AP/SRRH

FRAN handover procedures shift some control and decision-making toward edge devices and F-UEs. The paper reports lower handover signaling overhead in FRAN than in non-FRAN, conventional RAN, and CRAN.

  • FRAN handover procedures shift many handover-related functions, including admission control, from the gateway and BBU pool toward edge devices.Admission control occurs at the target MRRH; F-UEs can also take on some handover-related control and decision-making.Source: or in CRANs, the handover procedure… · p. 7
  • For F-AP-to-MRRH handovers, decisions are made between the source F-AP and F-AP gateway, and locally cached data need not pass through the core network.The procedure is also applicable to SRRH-to-MRRH handovers.Source: or in CRANs, the handover procedure… · p. 7
  • MRRH-to-F-AP handover is described as the most complex case because many target F-APs may be nearby and there is no direct communication about that handover.Source: or in CRANs, the handover procedure… · p. 7
  • For F-AP-to-F-AP handovers, decisions occur between the source F-AP and the F-AP gateway, while the two F-APs can communicate over the S1 interface.The handover procedure uses the traditional S1 and X2 interfaces; handover between two SRRHs is similar.Source: or in CRANs, the handover procedure… · pp. 7–8
  • Signaling overhead rises with session arrival rate and session holding time; the paper relates these trends to more handovers and increased cell-boundary crossing probability.The holding-time comparison is reported for λ = 0.1.Source: or in CRANs, the handover procedure… · pp. 10, 12, Figure 4 · p. 10
  • Handover signaling overhead is lower in FRAN than in non-FRAN for both F-AP-to-F-AP and F-AP-to-MRRH handovers.The paper attributes the reduction to local handover decisions and processing at F-UEs and F-APs, rather than relying on the BBU pool and core network.Source: or in CRANs, the handover procedure… · pp. 11–12

IV. INTERFERENCE MANAGEMENT AND RESOURCE OPTIMIZATION IN FRAN

The paper models uplink subchannel and power allocation for co-channel F-APs as a noncooperative game and proposes interference-aware resource allocation. Simulations report higher total net utility than existing FRAN and conventional LTE-A resource allocation.

  • The resource-allocation problem addresses co-channel interference, which the paper says can limit system throughput and energy efficiency.Dedicated F-AP channels are described as difficult, especially with many densely distributed F-APs.Source: §IV · p. 13
  • The scheme models uplink subchannel and power allocation for co-channel F-APs as a noncooperative game and uses interference-aware pricing to mitigate F-UE interference to the MRRH.Source: §IV · p. 13
  • Subchannel allocation maximizes each F-UE's net utility, combining achievable data rate, an interference price, and a reward for caching.Given the subchannel allocation, power control is modeled as a super-modular game; a Nash equilibrium is stated to exist on each individual subchannel.Source: §IV · p. 13
  • The simulated net-utility gain over non-FRAN grows with more F-UEs per F-AP, which the paper attributes to multi-user diversity; performance remains good as F-AP count increases.Source: §IV · p. 14
  • The proposed FRAN resource allocation performs better in total net utility than both existing FRAN resource allocation and conventional LTE-A resource allocation.Total net utility generally increases with the number of F-UEs per F-AP and with the number of F-APs.Source: §IV · p. 14

V. CONCLUSION AND FUTURE WORK

The article introduces an FRAN architecture for 5G networks, providing edge-of-network caching, computing, cooperative radio processing, and resource management.

V. CONCLUSION AND FUTURE WORK

  • The article introduces an FRAN architecture for 5G networks.Source: §V · p. 14
  • FRAN provides users with distributed local caching and computing.Source: §V · p. 14
  • FRAN supports cooperative radio signal processing and resource management at the network edge.Source: §V · p. 14

have proposed mobility management (including handover signaling procedures), interference

The proposed FRAN mobility and uplink resource mechanisms reduce handover signaling overhead and address cross-tier interference while increasing total net utility. The article also identifies security and integration challenges for future work.

have proposed mobility management (including handover signaling procedures), interference

functions for F-UEs. The Fog architectures should allow computing, storage, and networking

Future Fog architectures should support dynamic relocation of tasks among Fog, Cloud, and Things, with interfaces that enable interaction across those entities and users.

functions for F-UEs. The Fog architectures should allow computing, storage, and networking

Loading 1707.06892v1…