Elsevier

Computer Networks

Volume 164, 9 December 2019, 106907
Computer Networks

Analysis of a hybrid Fixed-Elastic DBA with guaranteed fronthaul delay in XG(s)-PONs

https://doi.org/10.1016/j.comnet.2019.106907Get rights and content

Abstract

Reusing the existing Optical Distribution Network (ODN) of Passive Optical Networks (PONs) to bring connectivity to residential households together with providing the so-called front-haul (FH) network in Cloud Radio Access Networks (C-RAN) is a promising converged fixed and mobile network architecture envisioned by network operators. This article overviews how to exploit Fixed Bandwidth Reservations allowed in ITU-T XG(S)-PONs and their suitability for the transmission of front-haul (FH) traffic, thus allowing one Remote Radio Unit (RRU) coexist with multiple residential Optical Network Units (ONUs) in the same PON. We explain how to configure such periodic reservations to improve access in the upstream channel of a PON, showing worst-case access delays below 100 μs for eCPRI’s functional split IU.

Introduction

In the last decade, fiber penetration has grown at a steady pace, reaching numbers of nearly 90% coverage and about 50% homes passed in the European Union (EU28) in September 2016 [1]. These numbers represent global averages per country, but in large cities fiber penetration is typically much higher. Indeed, Fiber to the Premises (FTTx) has been the technology of choice by a vast majority of network operators and telcos, given their benefits of cost-efficiency and high-performance metrics provided by current standards, namely ITU-T GPON [2], XG-PON [3], XGS-PON [4], NG-PON2 [5], and IEEE EPON [6], 10G-EPON [7] and the upcoming 100G-EPON [8]. The reader is referred to [9] for a detailed summary of current and future PON technologies.

In parallel, a trend towards deploying Cloud Radio Access Networks (C-RAN) [10], [11], where traditional Mobile Base Stations (BTS) are decoupled into low-complexity Remote Radio Units (RRUs) deployed in the streets and Baseband Units (BBUs) centralized (and possibly virtualized) in a cloud-like fashion, is gaining traction among the research community as a cost-effective architecture shift to be implemented and deployed in the fifth-generation mobile network (5G) [12].

In this light, a number of researchers have proposed to join both worlds on attempts to reuse the existing Optical Distribution Network (ODN) already deployed to bring connectivity to residential households together with providing the so-called front-haul (FH) network (the network connecting RRUs and BBUs) over the same underlying infrastructure [13], [14]. This strategy can provide important savings to network operators but, from a technology point of view, brings a number of important challenges that need to be investigated before actual deployment [15].

Depending on the functional split employed to divide the traditional BTS, essentially the number of functions left at the RRUs, FH traffic may require tens of Gb/s along with strict latency and synchronization requirements, as outlined in [16], [17]. If instead of using low-layer functional splits like CPRI, an upper functional split is used, the front-haul requirements may be relaxed allowing the coexistence of both residential traffic with packetized front-haul traffic.

Still, the upstream control protocol in TDM-based Passive Optical Networks (PONs), which typically employs a two-way Dynamic Bandwidth Allocation (DBA), needs to be carefully designed to favor the FH traffic offered by the RRUs at the expense of some penalty to residential traffic. Different Quality of Service (QoS) strategies for PONs have been studied in the literature to favor some Optical Network Units (ONUs) with respect to others. However, the strict needs of FH traffic require a careful design of the DBA strategy, focused on minimizing latency for RRUs.

This article presents a hybrid Fixed-Elastic Dynamic Bandwidth Allocation (FEDBA) algorithm that combines fixed bandwidth reservations to RRU-attached ONUs with elastic bandwidth reservations to residential households, showing that it is possible to re-use residential ODNs to serve FH traffic in certain scenarios. We show that some bandwidth over-provisioning to RRU-attached ONUs allows complying with the strict 250 μs delay limit of FH traffic without affecting excessively the traffic of residential ONUs.

In a nutshell, this article attempts to answer the following two questions:

  • 1.

    Can packetized FH traffic following eCPRI’s functional split IU be transported using TDM-PONs?

  • 2.

    If so, is the packet’s latency below the acceptable limits for such functional split?

The remainder of this work is organized as follows: Section 2 overviews previous work regarding the transport of FH traffic over Passive Optical Networks. Section 3 reviews important considerations regarding architectures and operational aspects of PONs. The proposed hybrid Fixed-Elastic DBA is detailed in Section 4. Some simulation experiments are carried out in Section 5. Finally, Section 6 gives the resulting conclusions.

Section snippets

Previous work

PONs have been deployed massively in the last decade, given their high bandwidth capacity (tens of Gb/s standard, hundreds of Gb/s envisioned), long reach (20 km standard, 40 km envisioned) and cost-effectiveness. Indeed, techno-economic studies suggest that operators amortize fiber deployments in a reasonable period of time while they provide sufficient capacity to meet the traffic demands of both residential and enterprise needs for the next decades [18], [19], [20]. Lately, the idea of

GPON, XGPON and XGS-PON review

Let us consider the cases of GPON, XG-PON and XGS-PON featuring either 2.5G/1.25G, 10G/2.5G and 10G/10G downstream/upstream capacity for connecting a number of residential ONUs and one or many RRUs, conceived to provide connectivity to both mobile and fixed services in scenarios with antennas on top of the roof of a residential building as shown in Fig. 1. The figure also shows the generic point-to-multipoint architecture of a generic PON showing its main architectural elements:

  • Feeder and

Overview of eCPRI’s split IU traffic profiles for 5G new radio

The 5G New Radio (NR) specified by the 3GPP initiative in Release 16 allows radio channels of up to 400 MHz with and Massive MIMO (up to 64 TX/RX) [43]. NR is based on OFDM very much like LTE but allowing a flexible numerology with subcarrier spacings ranging from 15 KHz up to 240 KHz [44]. In addition, Massive MIMO is well known to introduce important spectral efficiency gains, namely 8x2 MIMO provides 95% extra gain with respect to 2x2, while 16x2 and 64x2 MIMO increase the efficiency to 192%

Simulation experiments

In the following experiments, we have used a Discrete Event Simulator (DES) provided by the R-package simmer [46], [47] to simulate the different scenarios.

The residential ONUs are assumed to offer RONU=20 Mb/s traffic in a bursty nature, i.e. burst arrivals follow a compound Poisson traffic model where each ONU offers a burst of K packets following a Poisson distribution with a mean of E(K)=20 packets. Packet sizes are assumed to follow the classical empirically-observed AMS-IX (Amsterdam

Summary and conclusions

This paper proposes the use of periodic reservations of multiple overprovisioned (i.e. NT > 1 and α=7%) fixed transmission windows per 125 μs cycles to allow the transmission of eCPRI split IU fronthaul traffic over GPON, XG-PON and XGS-PONs. This strategy, in combination with a conventional Status Report based DBA can be used to enable the transmission of FH and BH traffic over a TDM-PON, giving rise to a hybrid Fixed Elastic DBA algorithm.

If properly dimensioned, we show via simulation that

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

The authors would like to acknowledge the support of the Spanish projects TEXEO (grant no. TEC2016-80339-R) and TAPIR-CM (grant no. P2018/TCS-4496). The authors would also like to acknowledge the support of the EU-funded project PASSION (grant no. 780326) to this work.

The authors would like to thank Dr. Iñaki Úcar for his support with the R-package simmer.

David Eugui completed his Bachelor’s degree in Telecommunications Engineering at Universidad Publica de Navarra (Navarra, Spain) in 2016, and theMaster in Telemat- ics Engineering at Universidad Carlos III de Madrid (Madrid, Spain) in 2018. Currently he is PhD Candidate at Univer- sidad Carlos III de Madrid.He became a teaching assistant in January 2019 and his research focuses on Machine Learning, Big Data, Computer Networks, Mobile Networks, Optical Networks. Mr. Eugui is part of the ADSCOM

References (50)

  • H.S. Abbas et al.

    The next generation of passive optical networks: a review

    J. Netw. Comput. Appl.

    (2016)
  • A. Pizzinat et al.

    Things you should know about fronthaul

    IEEE/OSA J. Lightwave Technol.

    (2015)
  • IDATE Digiworld Consulting, FTTH/B Panorama at September 2016, Available:...
  • ITU, ITU-t rec. g.984 gigabit-capable passive optical network, 2008, Available:...
  • ITU, ITU-t rec. g.987 10-gigabit-capable passive optical network (XG-PON) systems, 2012, Available:...
  • ITU, ITU-t rec. g.9807 10-gigabit-capable symmetric passive optical network (XGS-PON), 2016, Available:...
  • ITU, ITU-t rec. g.989 40-gigabit-capable passive optical networks 2 (NG-PON2), 2015, Available:...
  • IEEE, IEEE 802.3ah ethernet in the first mile, 2004, Available:...
  • IEEE, IEEE 802.3av-t rec. g.984 10 gbit/s ethernet passive optical network, 2009, Available:...
  • IEEE, IEEE p802.3cb working group 25 GB/S, 50 GB/S, and 100 GB/S ethernet passive optical networks task force, 2018,...
  • C. Mobile, The road towards green ran (white paper),...
  • A. Checko et al.

    Cloud RAN for mobile networks – a technology overview

    IEEE Commun. Surv. Tutor.

    (2015)
  • A. Checko et al.

    Evaluating c-RAN fronthaul functional splits in terms of network level energy and cost savings

    J. Commun. Netw.

    (2016)
  • T. Orphanoudakis et al.

    Exploiting PONs for mobile backhaul

    IEEE Commun. Mag.

    (2013)
  • D. Nesset

    PON Roadmap

    IEEE J. Opt. Commun. Netw.

    (2017)
  • P. Chanclou

    Mobile fronthaul architecture and technologies: a RAN equipment assessment

    IEEE/OSA J. Optical Comms. Netw.

    (2018)
  • 3GPP, “3GPP TR 38.801 V14.0.0 (2017-03): study on new radio access technology: Radio access architecture and...
  • A. De la Oliva et al.

    An overview of the CPRI specification and its application to c-RAN-based LTE scenarios

    IEEE Commun. Mag.

    (2016)
  • J.R. Schneir et al.

    Cost analysis of network sharing in FTTH/PONs

    IEEE Commun. Mag.

    (2014)
  • R. Sanchez et al.

    Provisioning 1 GB/S symmetrical services with next-generation passive optical network technologies

    IEEE Commun. Mag.

    (2016)
  • J.A. Hernández et al.

    Oversubscription dimensioning of next-generation PONs with different service levels

    IEEE Commun. Lett.

    (2016)
  • S. Kuwano et al.

    Operator perspective on next-generation optical access for future radio access

    Proceedings of the IEEE International Conference Communications

    (2014)
  • A.M. Mikaeil

    Performance evaluation of XG-PON based mobile front-haul transport in cloud-RAN architecture

    J. Optical Comms. Netw.

    (2017)
  • G. Kramer et al.

    IPACT A dynamic protocol for an ethernet PON (EPON)

    IEEE Commun. Mag.

    (2002)
  • F. Aurzada et al.

    Delay analysis of ethernet passive optical networks with gated service

    J. Opt. Commun. Netwo.

    (2008)
  • Cited by (6)

    • A hybrid dynamic bandwidth allocation scheme operating with IACG and cooperative DBA for converged fronthaul networks

      2023, Optical Fiber Technology
      Citation Excerpt :

      A number of DBA algorithms based on the ITU-T PON standard have been studied for use in the fronthaul. These include the Round Robin DBA (RR-DBA) [20], Group Assured GIANT (gGIANT) DBA [21], Optimized-Round Robin (Optimized-RR) DBA [22], Fixed-Elastic (FE-DBA) [23] and the Self-adjusting DBA [24]. In [20], a fixed maximum allocation bytes limit is used in its grant allocation which leads to inefficient utilization of the XG-PON upstream bandwidth [22].

    • Learning EPON delay models from data: A machine learning approach

      2021, Journal of Optical Communications and Networking
    • Performance of PON Dynamic Bandwidth Allocation Algorithm for Meeting xHaul Transport Requirements

      2021, 25th International Conference on Optical Network Design and Modelling, ONDM 2021
    • Learning from data: Applications of Machine Learning in optical network design and modeling

      2020, 2020 24th International Conference on Optical Network Design and Modeling, ONDM 2020

    David Eugui completed his Bachelor’s degree in Telecommunications Engineering at Universidad Publica de Navarra (Navarra, Spain) in 2016, and theMaster in Telemat- ics Engineering at Universidad Carlos III de Madrid (Madrid, Spain) in 2018. Currently he is PhD Candidate at Univer- sidad Carlos III de Madrid.He became a teaching assistant in January 2019 and his research focuses on Machine Learning, Big Data, Computer Networks, Mobile Networks, Optical Networks. Mr. Eugui is part of the ADSCOM research group and has participated in various research projects, both national and european. Some of these projects are: TIGRE5-CM, TAPIR-CM and PASSION.

    José Alberto Hernández completed the five-year degree in Telecommunications Engineering at Universidad Carlos III de Madrid (Madrid, Spain) in 2002, and the Ph.D. degree in Computer Science at Loughborough University (Leics, United Kingdom) in 2005. From 2005 to 2009, he was a postdoctoral researcher and teaching assistant at Universidad Autónoma de Madrid. In 2009, he moved to Universidad Carlos III de Madrid, where he became an associate professor. Dr. Hernández accumulates more than 10 years of teaching experience in areas related with Computer Networks, Computer Communications, Queueing Theroy and Optical WDM Networks; and has supervised 4 Ph.D. theses. Dr. Hernández has participated in a number of both national and european research projects concerning the modeling and performance evaluation of optical WDM networks, next-generation access networks, network energy efficiency, and optical transport for 5G mobile networks. Some of these projects are: e-Photon/ONe+, BONE, 5G-Crosshaul, DIOR, T2C2, CRAMnet, Medianet, TIGRE5-CM, etc. He has published more than 80 articles in both journals and conference in-proceedings on such subjects, including IEEE Communications Magazine, IEEE Network, IEEE Internet Computing, IEEE JSAC, Journal of Lightwave Technology, Journal of Optical Communications and Networking, etc. Dr. Hernández has also served as reviewer in many of the aforementioned journals.

    View full text