There's a huge choice of urban small cell backhaul technologies on offer today – and that's even before considering C-RAN fronthaul alternatives. We spoke with Greg Friesen, VP Products at Dragonwave, to get the perspective of a major established player in the microwave backhaul business.
C-RAN vs urban small cells
Today we are seeing more interest in small cell backhaul than C-RAN, but which one will win over the other can't easily be determined.. Almost everyone deploying C-RAN now is relying primarily on fibre and considering 70/80GHz E-Band point-to-point line-of-sight where wireless links are required. What's more, the physical size of the E-Band backhaul products is shrinking. For example, a transceiver with a 14cm dish can achieve 2.5Gbps, enough to support a single CPRI link.
The backhaul bandwidth for similar capacity urban small cells is much lower, typically 100 to 300Mbps. Co-located Carrier Wi-Fi could double those figures. Operators will want to support the upper end of LTE-Advanced capabilities, including Carrier Aggregation, with as little as 0.5ms or 1ms latency in the backhaul to ensure maximum spectral efficiency across a HetNet.
Outdoor urban small cell backhaul is still at a nascent stage. Operators are opting for fibre first, but recognise they will need wireless in due course. As a result, don't expect to see major take-up of wireless urban small cell backhaul during 2015 or perhaps even until mid 2016. Only about 5-10% of urban small cell backhaul installed in developed Asian countries is wireless at the moment.
A diverse set of products and variants
Urban small cells have many different use cases because regulations, spectrum availability, power supplies, and planning laws vary a lot between countries, and even municipalities, so there is no single "magic bullet" product that addresses every situation. There are so many unique requirements that it is inevitable a wide variety of products will be required. No single vendor can suddenly produce such diverse custom products in volume. This suggests that backhaul solutions won't be commoditised and that there will be many more options on the market than was the case for macrocells.
Operators will try to limit the range of tools in their toolkits, as the multi-national operators are already doing through standard equipment procurement specifications for cellsites. Even within the same country, achieving this is going to be difficult. For example, you might need a specific variant for New York that wouldn't work in Chicago, and therefore need to enlist two different vendors to satisfy both requirements. Add in factors such as power, colour and form factor, and one could foresee as many as 30 different variants for the same underlying product. To date, Dragonwave has prototyped many different concept variants for their customers to evaluate.
Urban Small Cells are primarily about capacity
There will be a few not-spots in the urban areas that require a small cell purely for coverage, but the vast majority will be installed for capacity. The backhaul architecture should not be a bottleneck and this drives a strong preference for overprovisioning where possible.
Wireless is quick to deploy and point-to-point 60GHz links would be a first choice, although they do require line-of-sight. Their high frequency reuse, relatively small form factor and minimal spectrum licence costs make them attractive. Proven products are available today with capacity of up to 1Gbps, and the sticker price has dropped by 50% in the last 2-3 years. Some of the earliest products are now in their third generation of advancement.
Point-to-Multipoint architectures are inherently more capacity constrained than point-to-point, but benefit from needing fewer transceivers. For maximum economic benefit, the hubs would ideally be able to connect with eight or more small cells, with a minimum of four. However, in most urban cases there is only Line-of-Sight available for one to three small cell deployments. This leads to either a Non-Line-of-Sight approach, or daisy chaining the backhaul between sites at street level.
Non-Line-of-Sight does have a place to play, but needs precious spectrum below 6GHz that is hard to come by. A good use would be to connect those small cells installed for coverage rather than capacity purposes.
The following table summarises these views:
|Point-to-Point|| || |
|Point-to-Multipoint|| || |
FDD vs TDD
Most 60GHz point-to-point links are FDD, taking full advantage of the wide spectrum bands available at that frequency.
Since TDD only requires a single slice of spectrum, it can be used in many more pockets of previously unused frequency bands. However, it does have an associated high delay of 2 to 8ms, as compared with 0.5ms for FDD. This can be a real killer when trying to meet the LTE specifications. While it may not rule out LTE-Advanced completely with a single link, a daisy chain of three or more may exceed the tolerance for packet delay variation.
Almost all NLoS products are TDD based today. Some use licenced TD-LTE spectrum, others use the unlicensed bands at 2.4 and 5GHz. In France, the 3.5GHz band is popular, but elsewhere this band is being considered for use directly for LTE access.
Speed of deployment is critical
The viability of urban small cell deployment depends on access to suitable sites and the speed at which they can be rolled out. By definition, their capacity will be required in the busiest metropolitan zones. So, avoiding the need to obtain special permits, or digging up or even closing down streets in order to lay down fibre, makes wireless small cells an attractive alternative. And, the costs for wireless are highly predictable for any given city. Fibre deployment costs vary drastically and can get out of hand if not carefully controlled.
For most products, you still need at least two people to deploy a link - one at each end. Mechanical alignment built into the product is still clunky and costly, and its reliability needs to be improved. In the long-term, electronic beam steering has more potential to be small and cost effective, with high, long-term reliability.
Even Non-Line-of-Sight products still need technicians at both ends during installation. These have wider beam widths and can bounce the signals off building walls. This makes it fairly easy to initialise the link, but achieving the maximum potential is much more difficult. Multi-path can result in a working link, but one that's been aligned via a sub-optimal path without realising it.
Their wide beam width and scattered signal also means that Point to Point NLoS links can sometimes interfere with each other, especially in dynamic street environments with moving traffic. Where spectrum permits, it's possible to select between different channels, and some systems have remote channel allocation and determination to enable this option.
- There's no "magic bullet" technology for urban small cell wireless backhaul; many different variants will co-exist
- 60GHz V-band point-to-point wireless links are the most broadly suitable for urban small cell backhaul, where fibre isn't easily available
- Capacity, not coverage, is the primarily driver for urban small cells. Backhaul should not be a bottleneck, implying preference for technical solutions which are not capacity constrained
- Rapid deployment that requires only minimal specialist training with simple, easy to use and maintain equipment is important to ensure a viable business case
Greg Friesen is VP Products at Dragonwave, a leading global supplier of wireless backhaul.
Read more about their small cell solutions here