Converged fixed-mobile access-metro optical network for 5G with programmable and elastic optical systems and SDN/NFV control
Raul Muñoz, Josep M. Fàbrega, Ricard Vilalta, Michela Svaluto Moreolo, Ramon Casellas, Vilalta, Laia Nadal, Ricardo Martínez, Centre Tecnològic de Telecomunicacions de Catalunya (CTTC/CERCA)
IEEE 5G Tech Focus: Volume 2, Number 2, May 2018
A holistic approach is essential in order to define a converged mobile and fixed access infrastructure, both at structural and functional level, instead of having multiple infrastructures delivering the same or similar services. This converged infrastructure should maximize: i) the benefits of the high-capacity and cost-effective optical access network solutions, addressing not only the transport requirements of fixed subscribers as done so far, but also for 5G mobile x-haul (fronthaul/backhaul) networks, and ii) the SDN/NFV architectural frameworks to address the challenge of jointly managing and operating heterogeneous access and transport networks and distributed cloud infrastructures to offer end-to-end network services and network slice services for fixed and mobile users.
Mobile radio access networks (RAN) and fixed optical access networks have evolved and developed independently. The aggregation and transport of mobile and fixed traffic in the access segment is implemented today by means of an overlay of heterogeneous transport network technologies tailored to some specific service. On the one hand, fixed optical access networks have been designed to meet the demands and requirements for the residential and enterprise subscribers (e.g. NG-PON1/2). These networks are typically passive optical networks (PONs), where a passive optical tree structure is used to connect the central office with the end users. The reason is that they are simple, flexible, and scalable, while featuring low maintenance and operation costs .
On the other hand, mobile transport networks can be traditionally sorted out as backhaul and fronthaul solutions. Mobile backhaul is the transport connection/network between the base-band units (BBUs) and the mobile core (i.e., EPC), and the mobile fronthaul is the transport connection/network between the remote radio head (RRH) and the BBU. Mobile backhaul has low requirements for bandwidth and delay, and traditionally is deployed in packet-based network infrastructure (e.g. Ethernet, IP/MPLS). Instead, mobile fronthaul has very stringent requirements in terms of high-bandwidth and low-delay because it is required to transport the digitally sampled radio waveform from the base station to the central office where the pool of BBUs is located. This architecture is known as cloud/centralized architecture (C-RAN), unlike the distributed RAN, where the BBUs are collocated with the RRHs in the base stations (e.g. 4G eNodeB). The most widely used standard interface for the fronthaul is the common public radio interface (CPRI) and is deployed in overlay C/DWDM links built specifically for this purpose, making use of optical transceivers supporting CPRI. These transceivers are known as digital radio over fiber (dRoF). The main drawback of CPRI is that it does not scale, in terms of bandwidth requirements, for the massive MIMO antenna deployments foreseen in 5G. 3GPP is proposing to reduce the bandwidth and latency requirements for 5G, while keeping most of the benefits of the C-RAN architectures, by performing a function split of the baseband processing in eight options that have been analyzed thoroughly in literature . In this approach, the BBU functions are split into the RRH and two new logical entities; central unit (CU) and distributed unit (DU). A CU can support multiple DUs, and each DU is only associated to only one CU. Most of the controlling functionalities are centralized on CU, while the fast scheduling on the air interface is realized on DU. It brings the introduction of two new transport networks between the CU and DU (known as midhaul), and between the DU and the RRU (known as next generation fronthaul interface – NGFI). 3GPP has already chosen split option 2 (i.e. split between PDCP and RLC) for the midhaul, but it is still open between the DU and RRH. This approach enables the virtualization of BBU functions (i.e., DU and CU) in local datacenters as well as the packetization of the NGFI to provide more efficient network utilization in ultra-dense scenarios. However, the packet-based NGFI will introduce new requirements in terms of jitter and synchronization that have to be properly addressed. An alternative solution that is being investigated to reduce the bandwidth requirements is to use analog radio over fibre (aRoF) transceivers, where the radio waveforms are directly modulated onto light for connecting BBUs and RRHs.
Additionally, the wide adoption of Network Function Virtualization (NFV) concepts, including BBU virtualization requires cloud services for the deployment of virtualized network functions (VNFs). The virtualization of network functions that are typically deployed in specialized and dedicated hardware (e.g. mobile evolved packet core –EPC) is of crucial importance for 5G. Much like 5G, IoT also requires core computing and storage infrastructures in order to perform IoT analytics from the data collected from sensors and actuators (e.g., temperature monitoring, energy consumption measurement, etc.). Traditionally, cloud services have been implemented in large datacenters (DCs) in the core network. Cloud offers high-computational capacity with moderate response time, meeting the requirements of centralized services with low-delay demands. However, there is a general trend both at mobile and IoT level to offer computing services at the edge of the network leveraging on ultra/low-latency and high-bandwidth. For example, ETSI is defining the mobile edge computing (MEC) to offer applications such as video analytics, location services, mission-critical applications, augmented reality, optimized local content distribution and data caching, that can be considered as VNFs. Thus, it is required to dynamically allocate computing and storage resources to flexibly deploy VNFs in multiple DCs, and to provide the required connectivity between DCs with Quality of Service (QoS). Additionally, another main requirement is to offer network slices existing in parallel and isolated for different tenants (e.g., vertical industries, virtual operators) in order to deliver the tenant-specific requirements (e.g, security, latency, resiliency, bandwidth) .
To meet the above transport and service requirements, this paper presents a converged fixed-mobile access-metro optical network with programmable and elastic optical systems and distributed datacenters for cloud and edge computing. Additionally, this paper presents a service platform to provide efficient end-to-end resource and service orchestration based on reference architectures such as the ONF Software Defined Networking (SDN) and ETSI NFV standards.
Figure 1: Converged fixed-mobile access-metro optical network with unified SDN/NFV control framework
2. Elastic optical access and transport
Fig. 1 shows the scheme of the network concept proposed. There, the COs are connected to the edge nodes (ENs), where BSs are located, by means of a passive optical network (PON) scheme. In order to deliver fixed access services to the users, the CO hosts the corresponding optical line terminals (OLTs) and aggregation subsystems. Interestingly, selected mobile x-haul signals can be multiplexed/demultiplexed in wavelength at the CO and transparently routed to another node of the metro/core network for further processing . A PON scheme is envisioned as external plant in order to leverage the existing optical access infrastructure (e.g. GPON, NGPON1). By following this approach, the entire C-band is available in the majority of deployments. Therefore, a wavelength overlay of channels can be envisioned for providing C-RAN services over the exiting fixed access infrastructure (shared with fixed users) while following the elastic networking paradigm . Additionally, when following the specifications of NGPON2 , virtual point-to-point links can be established by means of wavelength division multiplexing (WDM). These channels could be assigned to different ENs and/or services. Nevertheless, NGPON2 envisions also a high-speed residential access service employing time and wavelength division multiplexing (TWDM), where part of the C-band is used for downstream. Therefore, a careful analysis should be performed for the specific deployments of NGPON2. In terms of network architecture (shown in Fig. 1), a common power-splitting tree can be envisioned as a general case. In case of pursuing NGPON2 compatibility, different options can be envisioned. An interesting example can be including an intermediate WDM distribution stage in order to implement a hybrid wavelength-switched/-routed PON (as shown in Fig.1).
Additionally, the proposed network architecture can envision the gradual upgrade of some network parts (e.g. feeder) in order to include spatial division multiplexing (SDM) for further increasing the network capacity. Therefore, the spatial diversity obtained, in combination with WDM, would enable a fronthaul/backhaul infrastructure with unique 2-dimensional (2D) properties. A first approach for SDM can be based on bundles of standard single-mode fibers (SSMFs), since cables deployed in the field typically have a loose-tube design containing several fibers . Nevertheless, a longer-term solution can rely on multicore fibers (MCFs), providing a compact parallel transmission medium.
By assuming the introduction of a flexible functional split, bandwidth and latency requirements are reduced, enabling the use of statistical multplexing. Therefore, cost-effective Ethernet can be employed to provide flexible allocation of capacity to the high number of endpoints and users in ultra-dense scenarios connected through C-RAN and fixed technologies. In order to allow aggregation and switching of flows including quality of service (QoS) at the transport level, the COs and ENs are expanded with Carrier Ethernet switches.
At the CO/OLT, programmable sliceable bandwidth/bitrate variable transceivers (S-BVTs) are used for concurrently serving different ENs at variable capacity. These S-BVTs can deliver multiple flows/slices where each can be independently configured by the control plane. At the other end of the network, each EN has a programmable BVT (non-sliceable) at the optical network terminal (ONT). The (S-)BVTs can be remotely configured by the control plane, for an optimal management of the network resources . The parameters to be configured at each (S-)BVT include wavelength, spectral occupancy and capacity per flow. So, the (S-)BVTs deliver data flows with variable spectral occupancy and rate, according to the network and path conditions. Consequently, this solution is specifically tailored mobile midhaul, enabling the optimal management of different functional splits, while being also well-fitted for mobile fronthaul/backhaul.
Among all the options for implementing the (S-)BVTs, those based on DD orthogonal frequency division multiplexing (DD-OFDM) are the most attractive. In fact, these transceivers can be ad hoc configured for achieving a certain reach and/or coping with a targeted data rate adopting low complex optoelectronic subsystems .
3. Multi-tenant SDN/NFV control and orchestration
The considered SDN/NFV control and orchestration system provides NFV network services and network slicing services. It is deployed on top of the end-to-end infrastructure shown in Fig. 1, composed of multi-layer (packet/optical) networks and multiple NFV infrastructure point of presence (NFVI-PoP) at the edge and core of the network. An NFVI-PoP is a set of computing, storage and network resources that provides processing, storage and connectivity to VNFs through the virtualization layer (e.g. hypervisor). It is deployed in micro-DCs in the edge nodes, small-DCs in the COs, and core-DCs in the core network. The ETSI NFV management and orchestration (MANO) architectural framework  identifies three functional blocks; virtualized infrastructure manager / WAN infrastructure manager (VIM/WIM), NFV orchestrator (NFVO) and VNF manager (VNFM).
The VIM is responsible for controlling and managing the NFVI-PoP’s virtualized compute, storage and networking resources, whilst the WIM is used to establish connectivity between NFVI-PoP’s. The VIM is commonly implemented using a cloud controller based on OpenStack. It interfaces with the NFVO/VNFM reference implementations using the OpenStack API. OpenStack enables to segregate the resources into availability zones for different tenants and to instantiate the creation/ migration/ deletion of VMs and CTs (computing service), storage of disk images (image service), and the management of the VM/CT’s network interfaces and network connectivity (networking service). The WIM can be performed by dedicated Transport SDN controllers (e.g. OpenDaylight, ONOS, Ryu) in charge of managing the packet and optical technologies. However, the main limitation of this approach is that currently the interface between the NFVO and the WIM is not widely implemented and still lacking maturity 
The VNFM is responsible for the lifecycle management of VNF instances, and the NFVO has two main responsibilities; the orchestration of NFV infrastructure resources across multiple VIMs and WIMs (resource orchestration) and the lifecycle management of network services (network service orchestration). The network service orchestration is responsible to coordinate groups of VNF instances that jointly realize a more complex function (e.g. service function chaining), including joint instantiation and configuration of VNFs and the required connections between different VNFs . The NFVO and VNFMs are typically implemented together in reference software implementations such as open source MANO (OSM), ONAP or SONATA.
Finally, a Network Slice Manager (NSM) is deployed on top of the NFVO. The NSM is responsible of the lifecycle management of the network slices. Network slicing extends related concepts such as ETSI network services by defining a network slice instance as one network service instance or a concatenation of network service instances. The NSM leverages the NFV MANO to provide the network services associated to a slice and fulfilling its deployment requirements (e.g, security, latency, resiliency, bandwidth). The MSM is not supported in the current NFV MANO implementations, and it relies on proprietary extensions.
5G is targeting a converged x-haul (fronthaul/midhaul/backhaul) network and cloud infrastructure to offer end-to-end services. To this end, it is required to move from the traditional overlay of heterogeneous and independent networks to a novel architecture integrating both mobile and fixed transport networks. This paper has presented a converged fixed-mobile access-metro optical network by integrating the recent advances in flexible and programmable optical device technologies, together with the emerging SDN/NFV control and orchestration paradigm.
Work supported by EC H2020 BLUESPACE (762055) and the Spanish DESTELLO (TEC2015-69256-R) projects.
- A. Girard, FTTxPon technology and testing, EXFO Electro-Optical Engineering Inc, 2005
- Chih-Lin I, Han Li, Jouni Korhonen, Jinri Huang, Jinri Huang, RAN Revolution with NGFI (xhaul) for 5G, Journal of Lightwave Technology, DOI: 10.1109/JLT.2017.2764924
- 5GPPP white paper, the 5G Infrastructure Public Private Partnership: the next generation of communication networks and services, March 2015.
- J. M. Fabrega et al., “Experimental Validation of a Converged Metro Architecture for Transparent Mobile Front-/Back-Haul Traffic Delivery using SDN-enabled Sliceable Bitrate Variable Transceivers,” in Proc. of ECOC 2017, paper M.2.A.5
- D. Nesset, "NG-PON2 Technology and Standards," in Journal of Lightwave Technology, vol. 33, no. 5, pp. 1136-1143, Mar. 2015.
- M. Svaluto Moreolo et al. “SDN-Enabled Sliceable BVT Based on Multicarrier Technology for Multiflow Rate/Distance and Grid Adaptation,”in Journal of Lightwave Technology, vol. 34, no. 6, pp. 1516-1522, Feb. 2016
- Network Functions Virtualisation (NFV); Management and Orchestration, ETSI GS NFV-MAN 001 v.1.1.1, (2014-12).
- Raul Muñoz, Ricard Vilalta, Ramon Casellas, Arturo Mayoral, Ricardo Martínez, Integrating Optical Transport Network Testbeds and Cloud Platforms to Enable End-to-End 5G and IoT Services, in Proc. of ICTON 2017.
- R. Casellas, R. Muñoz, R. Vilalta, R. Martínez, Orchestration of IT/Cloud and Networks: From Inter-DC Interconnection to SDN/NFV 5G Services , in Proceedings of the Optical Networks Design and Modelling (ONDM2016) conference, May 2016.
Raul Muñoz (SM’12) graduated in Telecommunications Engineering in 2001 and received a Ph.D. degree in Telecommunications in 2005, both from the Universitat Politècnica de Catalunya (UPC), Spain. Currently, he is Head of the Optical Network and System Department. Since 2000, he has participated in over 35 R&D projects funded by the Spanish and EC’s Framework Programmes (H2020, FP7, FP6 and FP5), as well as industrial contracts. He has been Project Coordinator of 5 Spanish projects, the coordinated EU-Japan FP7-ICT STRAUSS project (608528), and the H2020-MSCA-ITN ONFIRE project (765275). He has published over 60 journal papers and 200 international conference papers.
Josep M. Fabrega (S’05-M’10-SM’17) received his BSc, MSc and PhD degrees in telecommunications engineering from UPC-BarcelonaTech, Barcelona, Spain, in 2002, 2006 and 2010, respectively. Currently he is a Senior Researcher in the Optical Networks and Systems Department of CTTC, Castelldefels, Spain. Since 2004, he has been actively involved in 22 public and industrial projects. He is the author/co-author of more than 100 papers, including 2 patents. His research interests include optical communication systems and signal delivery over novel optical network architectures. Dr. Fabrega is an IEEE Senior Member and received the EuroFOS best student research award in 2010.
Ricard Vilalta (M’13, SM’17) graduated in telecommunications engineering in 2007 and received a Ph.D. degree in telecommunications in 2013, both from the Universitat Politècnica de Catalunya (UPC), Spain. Ricard Vilalta is senior researcher at CTTC, in the Optical Networks and Systems Department. He is an active member of ETSI, IETF and ONF standardization bodies. His research is focused on SDN/NFV, Network Virtualization and Network Orchestration. He has been involved in several international, EU, national and industrial research projects. He has also authored and co-authored more than 180 journals, conference papers and invited talks.
Michela Svaluto Moreolo (S'04-A'08-SM'13) received the M.Sc. degree in Electronics Engineering and the Ph.D. degree in Telecommunications Engineering from University Roma Tre, Rome, Italy, in 2003 and 2007, respectively. She currently is a Senior Researcher and the Coordinator of the Optical Transmission and Subsystems research line within the Optical Networks and Systems Department, at the Centre Tecnològic de Telecomunicacions de Catalunya (CTTC), Castelldefels, Spain. She also serves as member of the CTTC Management Team, with the role of Project Management Coordinator. Her research interest areas include advanced transmission technologies and software-defined systems for future optical networks.
Ramon Casellas (M'09, SM'12) graduated in telecommunications engineering in by the UPC-BarcelonaTech and ENST Telecom Paristech (1999). He worked as an undergraduate researcher at France Telecom R&D and British Telecom Labs, and completed a Ph.D. in 2002 at ENST, working as an associate professor. He joined CTTC in 2006, working in international and technology transfer research projects. His research interests include network control and management, traffic engineering, GMPLS/PCE, SDN and NFV aapers, 4 IETF RFCs and 4 book chapters.
Laia Nadal (S’10) obtained her MSc Telecommunication Engineering degree July 2010 and the Master of Research on Information and communication Technologies in 2012. She received her PhD degree in November 2014. In 2010, she was awarded with the FPI grant from the Spanish Ministry of Economy and Competitiveness (MINECO) to perform her PhD in the CTTC. From May to August 2013 she was a Visiting Ph.D. Scholar in ADVA Optical Networking (Germany). She currently is a Researcher of the Communication Networks Division. Her research interests include signal processing and advance modulation formats for optical communication systems.
Ricardo Martínez (SM’14) received an M.Sc. degree in 2002 and a Ph.D. degree in 2007, both in telecommunications engineering, from the UPC–BarcelonaTech University, Spain. He has been actively involved in several EU public-funded and industrial technology transfer projects. Since 2013, he is Senior Researcher in the Communication Networks Division at CTTC in Castelldefels, Spain. His research interests include control and orchestration architectures for heterogeneous and integrated network and cloud infrastructures along with advanced mechanisms and algorithms for provisioning and recovering of quality-enabled services.
Editor: Rod Waterhouse
Subscribe to Tech Focus
Join our free IEEE Future Networks Technical Community and receive IEEE Future NetworksTech Focus.
Article Contributions Welcome
If you wish to have an article considered for publication, please contact Amine Maaref, Managing Editor, at firstname.lastname@example.org or Siming Zhang, Associate Managing Editor at email@example.com.
Geoffrey Li, Editor-in-Chief
Amine Maaref, Managing Editor
Siming Zhang, Assoc. Managing Editor
Imran Shafique Ansari
Zhi Ning Chen