The backhaul challenge
25 May 2017 |
5G services will remain on the distant horizon until backhaul networks are fit for purpose, but how can carriers deliver what’s needed and keep costs under control, asks Sue Tabbitt
The Institute of Electrical and Electronics Engineers (IEEE) set out its concerns in a study, 5G Backhaul Challenges and Emerging Research Directions: A Survey.
This explored the practical challenges that need to be addressed as the telecoms industry aligns itself to take advantage of 5G – in particular, the potential backhaul bottleneck.
As ultra-dense and heavy traffic cells try to connect to the core network, with substantial requirements around capacity, latency, availability, energy and cost efficiency, the vision for innovative new services based on the IoT and beyond will be hard to realise unless backhaul networks are fit for purpose.
The IEEE report, published in 2016, concludes that no single solution can solve the overall 5G backhaul problem. It advocates an optimised backhaul
- that is dynamic, adaptive and self-organising
- which blends the best of existing transport network architecture with new technologies;
- where radio access networks are fused with the backhaul – they are jointly designed, operated and optimised;
- where IT techniques such as SDN are instrumental.
Hanna Maurer Sibley, head of network products for central Europe at Ericsson, agrees that there is little point worrying about 5G spectrum until the backhaul is ready. This includes meeting the need for radio base station synchronisation, to minimise interference once uplinks and downlinks are on the same spectrum.
“It’s easy to just talk about the radio front-end, and of course operators are traditionally organised by radio and transport, but to make 5G happen the two ends need to talk to each other,” she says. Ericsson can help with that, she says, because it caters for both in its technologies and services.
“If my salary depended on it, I’d be focusing on synchronisation, latency and security,” she says of where wholesale infrastructure providers should be focusing their efforts currently – for example to support the kinds of applications 5G will be used for – such as connected cars, smart cities and wider IoT deployments.
Beyond highly populated urban areas, likely use cases could mean targeting location-based investments in areas close to transport networks.“Germany is very strong on automotive, so is an obvious market for connected cars, so it makes sense to ensure infrastructure next to highways,” Maurer Sibley says. Train networks could be another focal point, as well as major transport and logistics hubs such as busy shipping ports. “We’ve just helped Rotterdam harbour put in a private LTE network,” she says.
In developing regions, opportunities may be more linked to solving practical issues – connecting the unconnected and getting services out to hard-to-reach areas. Investment here is likely to be guided by government priorities and grants.
Satellite connections have traditionally provided the only real answer to getting connections to hard-to-reach regions, though hardly the cheapest or highest-performing. Issues beyond unsustainable costs include high latency and bottlenecks, notes Carsten Brinkschulte, CEO of Core Network Dynamics.
“We need to take the strain off the satellite network,” he says, pointing to virtualisation and mobile edge computing (MEC – now also known as multi-access edge computing) as a more viable alternative. This offers an IT service environment with cloud-computing capabilities at the edge of the network, enabling decentralised core networks which provide connectivity and computing to the edge, running applications and even the evolved packet core (EPC) close to devices.
“By using the satellite network for off-net traffic only, this approach keeps local voice and data communications local, avoiding backhaul wherever possible and offers greater resilience and more reliable connectivity,” Brinkschulte says.
Alternative deliveries by satellite
Another option is to combine broadband satellite services with small cells. This is a scenario supported by satellite services provider Hughes Network Systems, which is also a member of the Small Cell Forum. Targeting specific areas, rather than extending mobile networks indiscriminately, can help to make the cost versus returns add up – especially in the Middle East, Africa and Asia and Latin America.
Hughes supports this kind of model via its Jupiter satellite backhaul system, which provides dynamic bandwidth allocation matched to cell site needs. It has now enhanced the Jupiter system with 4G LTE acceleration, enabling mobile operators to deploy fast mobile services anywhere in the world, says the company.
In Bolivia, state-owned fixed and mobile operator Entel has already signed up to use Hughes’s platform at more than 500 deployed sites to extend cellular provision, including LTE access and internet access, to remote areas of the country.
The harder they save
The harder case for carriers to make is for capex investment in new 5G fibre networks. Ceragon Networks, a specialist in high-capacity microwave Ethernet and TDM wireless backhaul, recently announced that a tier-one mobile operator in India had already placed over $60 million in new orders so far this year, in a further wave of wireless backhaul deployments using its IP-20 platform.
The operator wants to connect 4G LTE sites and deliver subscriber-based HD multimedia services across India.
“While fibre backhaul undoubtedly makes sense for some use cases, total cost of ownership is much higher than wireless backhaul technology in use today and, in terms of practicality, particularly where the IoT is concerned, it doesn’t make monetary or logistical sense,” says Ravi Palepu, global head of telco solutions at VirtusaPolaris, a global IT services company. “Carriers must look to a combination of wireless backhaul options as future-proof and scalable alternatives.”
Yes, investment will be needed to build the greater numbers of base stations that 5G will call for, he says, but it’s likely that most will run point-to-multipoint connections, allowing carriers to aggregate backhaul traffic from numerous base stations at a fraction of the cost of fibre.
He cites the potential of small cells, which will increasingly become an integral part of wireless networks, providing additional pockets of capacity as needed.
“By appearing in non-traditional locations such as utility poles or street lamps, or on the sides of buildings, carriers can avoid many location and other charges too,” Palepu says. “With this added capacity, plus the savings made from avoiding costly fibre projects, wireless backhaul will remain the most cost-effective and pragmatic approach to managing 5G’s massive data demands.”
Self-organising and self-sealing
Cambridge Communication Systems (CCS), a multipoint-to-multipoint wireless backhaul and transport systems provider, offers a self-organising microwave backhaul system – Metnet.
FiberTower in the US, which has just been acquired by AT&T, recently announced plans to use this to provide a wireless backhaul, transport and 5G infrastructure solution with line-of-sight and non-line-of-sight connectivity. The carrier owns wide-area authorisations in the 24GHz and 39GHz bands. The microwave backhaul system is FCC-certified for operation in the 24GHz spectrum band, where application opportunities include small cell and fixed wireless broadband backhaul, and fibre extension.
Its nodes connect autonomously to form self-organising, self-healing links that dynamically reconfigure to optimise performance and spectral efficiency as line-of-sight circumstances or traffic levels change. This happens while still minimising interference.
So what of the need for the front-haul and backhaul providers to come together to smooth the way for 5G? It begins with thinking differently.
The new way of thinking
“Carriers are turning to new ways of investing to cut down costs,” says Shany Elkarat, director of product and technology at Colt. “First they are beginning to plan more around long-term investments when looking for IRUs [indefeasible rights of use] over 15 years to get fibre in the ground.” Carriers are also moving to a shared backhaul approach such as a cloud radio access network and shared infrastructure, rather than sole investments. Moving assets and infrastructure to neutral data centres is a further option, aiding cost-effectiveness and flexibility, he says.
One option for backhaul optimisation, according to Jon Baldry, head of marketing for Infinera’s metro business unit, is packet-optical solutions. “This can remove the need for some dedicated Layer 3 IP devices within the network, while also improving the transport performance of the network regarding latency and synchronisation support, which will be essential in 5G networks,” he says.
Supporting increased bandwidth
“These networks can help address the economic pressures on backhaul links and lay the foundations for software defined networks, and future moves to networks functions virtualisation.” Operators don’t need to wait for 5G to evaluate these approaches to mobile transport, he says.
“Supporting the increased bandwidth of 5G will require wide-scale network evolution,” says Lionel Chmilewsky, CEO of CBNL. He claims the company’s point-to-multipoint technology enables carriers to deploy a licensed backhaul or fixed wireless network in half the time, and with half the TCO compared to more traditional point-to-point technology.
He says that adding this technology is already driving some of the largest and most modern backhaul networks in regions such as the Middle East and Africa, where CBNL’s customers include MTN, Airtel, Vodacom and Mobily.
Wherever they are, carriers don’t and shouldn’t wait for 5G to deploy high-capacity services, he says. “They should act now, or they run the risk of risk being left behind.”