Quantifying the cost benefits of FTTH for 5G transport
A Bell Labs Consulting study
5G is set to change the world. It is at the heart of every operator’s strategy, but will require significant investment. To be successful, operators need to find ways to cost-efficiently deploy 5G. PON brings that efficiency, and high-performance fiber access networks will play a key role in delivering 5G experiences everywhere. PON supports both the performance demands (throughput and latency) and RAN densification that 5G requires, and FTTH networks are already deployed where 5G will be in most demand, namely dense urban and semi-urban environments.
A recent analysis by Bell Labs Consulting shows that leveraging existing FTTH networks can decrease 5G transport costs by more than 50% compared to traditional solutions (microwave, P2P fiber). In this paper we present the assumptions, considerations and results of this study.
5G anyhaul scenarios
The Bell Labs study compares the total cost of ownership (TCO) over a 5-year period for various mobile transport (anyhaul) technologies.
The starting assumptions are that the mobile network operator (MNO) wants to densify the RAN through the introduction of small cells in an area where FTTH is already deployed and has sufficient spare fiber assets that can be reused for mobile transport.
We consider the small cell and an MNO hub site as the two endpoints for the transport domain, and also consider backhaul and midhaul scenarios, i.e. the connectivity between the central unit (CU) and distributed unit (DU).
The study compares the following transport solutions:
- Dark point-to-point (P2P) fiber: every small cell (SC) has a dedicated P2P connection to a designated MNO macro cell hub.
- Hybrid microwave transport (D- or E-band) and P2P fiber: the traffic from a cluster of neighboring SCs is aggregated via one or two chains of mmWave links, and only one P2P fiber path is required to connect a SC cluster with the MNO macro cell hub.
- Dark fiber PON: the MNO leases a dark fiber PON and operates active PON equipment. Access nodes (Optical Line Termination, OLT) are positioned at the macro cell hub site and Optical Network Terminations (ONT) at the small cell.
- PON capacity lease on a dedicated PON: the MNO leases capacity on a dedicated PON for mobile transport from a FTTH Infrastructure Provider (FTTH InP) who operates the OLTs. A P2P link is established to exchange the aggregated traffic between the CO and the MNO hub.
- Capacity lease on overlay PON: the MNO leases capacity on a PON that is also used for residential, business users, and small cell transport. Residential and small-medium businesses are served on a GPON wavelength, and mobile transport and large businesses on XGS-PON or 25G PON wavelength in overlay.
- Converged operator: the MNO also owns the FTTH network, using a dedicated PON for transport (similar to case 4).
Figure 1: Mobile transport architectures
Table 1: Overview of mobile transport solutions
The analysis shows the factors that most influence TCO are:
- Fiber lease cost.
- Network architecture (physical collocation of CO and MNO hub).
- Number of small cells that can be multiplexed on a shared link (PON or microwave chain).
A large TCO contributor is the annual lease cost charged by the FTTH InP. This depends on many factors, such as the number of fiber assets (P2P based anyhaul will need more fibers than point-to-multipoint PON), and fiber length (scarce and long feeder fibers versus the drop fibers with typically a much smaller strand length). PON point-to-multipoint (P2MP) saves on the fiber count in the feeder section and space at the fiber concentration point, resulting in lower OPEX and lower lease costs.
Obviously, the MNO and the FTTH InP will have to negotiate the lease contract specifics (Service Level Agreement, pricing, etc.). If the wholesale price is too high, the MNO may decide to build its own strategic PON network with additional capacity. In this way, the MNO will have its own transport network and can expand its service offering to residential and business customers, creating new revenues and competition for FTTH providers.
The study considers that the total number of small cells is the same for all transport cases. However, for P2MP, there is a difference in how many cells can be grouped per PON or microwave cluster, ultimately impacting the cost (number of fiber assets, active equipment footprint, etc.).
Capacity is one of the factors. The number of small cells that can be statistically multiplexed on a shared cluster/link is determined by RAN peak capacity, multiplexing gain and capacity of the transport medium and technology. Calculations show that PON capacity is sufficient to enable tens of small cells per PON, which is more than enough for backhaul/midhaul in urban areas.
Other constraints include geography and line-of-sight (LoS). With PON overlay transport, even if the PON capacity can support a higher number of small cells, in reality there may not be so many cells to connect. In urban and semi-urban areas, the small cell density will be 10-100 times lower than FTTH subscribers, which means typically 2-3 small cells connected per PON. For LoS constraints in microwave transport, we assume a maximum of 7 DUs (midhaul) in a cluster/chain.
The MNO hub site will in general not coincide with the CO, where fiber feeders are terminated and OLTs are typically located. In most cases, the connectivity between two anyhaul endpoints will be based on a two-legged (two hops) route via the CO. One-hop solution, with a direct fiber link between hub and small cells, is rather an exceptional case.
- With a dark fiber/dark PON lease, fiber strands towards the small cells and towards the hub are connected at the optical distribution frame (ODF) in the CO.
- With a PON capacity lease, an OLT located in the CO aggregates the traffic from multiple SCs on multiple PONs, and a high capacity P2P link funnels the aggregated traffic to the hub.
Why is this important? Our analysis shows that a one-hop versus a two-hop approach makes little difference to TCO for PON based transport (5-7%). However, the impact on P2P/ P2P + microwave is significant which can make these solutions more attractive than PON in case of single-hop architecture but only in markets with higher lease prices.
Figure 2 compares the midhaul access transport TCO (per small cell) at a low fiber lease price level for the two-hop scenario where traffic of up to 6 DUs (x 2 RUs per DU) can be multiplexed on a XGS-PON or E-band microwave cluster.
Figure 2: Backhaul/midhaul 5-year TCO comparison
- PON-based transport is the most cost efficient, even with high fiber lease prices and irrespective of the number of small cells (as long as they can be supported by the PON capacity).mm
- Compared to a new build of a P2P fiber network, PON transport is about 10-times less expensive.
- The biggest advantage is for converged operators that own a FTTH network. The cost advantage can go beyond 60% compared to non-PON based transport. In this case, the biggest cost is associated with building out the last 10-20m from the fiber distribution point to the cell itself. A much smaller part of the cost is allocated to active equipment (OLT PON ports and ONTs at the cell site).
- For MNOs that don’t own FTTH, the TCO advantage is approximately 50% compared to P2P dark fiber lease + microwave, and 40% compared to P2P dark fiber lease. The major cost contributor is the bitstream lease cost. There is a small cost for active equipment, like ONTs at the cell site, because we assume the mobile operator would prefer to connect the SFP ONTs to cells, rather than relying on a fixed operator.
- The PON capacity lease on a dedicated PON is more cost-efficient than PON capacity lease in overlay, for scenarios with higher small cell density. This is because the overlay scenario uses the existing residential FTTH PON and can have 2-3 small cells per pre-established FTTH PON. For dedicated PONs, where the PON outside plant is designed and built to perfectly match small cell locations, the number of small cells is only constrained by the PON capacity.
- microwave + P2P fiber becomes more attractive for scenarios with higher fiber lease prices (for example, in markets with scarce fiber) or in a single-hop approach.
PON-based transport is the most cost-efficient way of addressing small cell transport. A clear advantage for MNOs when leasing FTTH fiber assets is the immediate availability of the anyhaul service and the fact that no large upfront investment is required, compared to an own-build solution. Leasing could have different forms: for example, a capacity lease (corresponding with bit stream) on a shared or dedicated PON, or a dark fiber lease. Next generation PON technologies (XGS-PON and 25G PON) ensure that ever-increasing capacity demands of transport networks can be met on the same fiber infrastructure.
For FTTH infrastructure providers, the anyhaul service ensures a steady extra revenue income, accelerating fiber network monetization. FTTH providers should consider supercharging their fiber networks and making them attractive for mobile transport in a 5G world.
Thanks to Bell Labs Consulting experts Tom Van Caenegem and Vincent Bonnet for the insights and research provided.
Share your thoughts on this topic by joining the Twitter discussion with @nokia or @nokianetworks using #anyhaul #5G #Broadband