Backhaul Evolution - The elephant in the room
Clearwire's CTO says it forms the highest cost of a network deployment, so what 
solutions provide the best answers for the future of mobile backhaul, asks Alan Solheim.
The evolution from voice only cellular systems to 3G+ (HSPA, WiMAX, LTE...) has revolutionized the way we interact, share information, work and play. In order to deliver on this promise, the entire network - from the handheld device to the core of the network - has to change at every level. One area that has not only changed, but has been turned inside out, is backhaul. Long considered only a cost of doing business, many carriers are starting to see backhaul as an enabler for new service delivery, and in some instances like ClearWire in North America, a competitive advantage over traditional players. ClearWire has built the largest greenfield WiMAX network to date and their CTO, John Saw says, "It's what I call the elephant in the room that nobody talks about. Backhaul is probably the highest cost in deploying a network. Anyone who wants to roll out a real wireless broadband network nationwide needs a cheaper solution than current models." This same sentiment is true for anyone planning to build a 3G+ network: without radical change to the backhaul, the applications will be starved for bandwidth, the user experience will be unacceptable, and the network economics will not be favorable.
So what are the required changes? The most obvious one is in bandwidth. 2G voice networks need a single E1 connection to the base station in order to provide the required capacity. The advent of GPRS and EDGE to provide data services, resulted in an increase of up to 4 E1s per base station, but leased circuits or low capacity TDM microwave radios were able to provide the increased capacity. The introduction of High Speed Packet Access (HSPA) and HSPA+, has driven the capacity per base station up by a factor of 10, straining the throughput capability of these types of microwave and making leased E1 circuits cost prohibitive.
As 3G base stations began to support native Ethernet interfaces, enabling the use of packet microwave or leased Ethernet backhaul, a variety of approaches were adopted in order to support legacy E1 interfaces. One method has been to leave the legacy E1 transport for the 2G base stations intact while adding an overlay to support new HSPA/HSPA+ base stations. This has been more prevalent among operators who have used leased E1 circuits for the 2G backhaul. Alternative deployments have included hybrid TDM/Ethernet microwave, or packet microwave with pseudowire. Finally, fibre has generally been used when available at the cell site however fibre penetration is very low - even in developed countries.
With the advent of 4G technologies (WiMAX and LTE), the network is IP end to end, and the backhaul load per base station has again gone up by almost another order of magnitude. Furthermore, in order to deliver the desired user experience the base station density has to increase: between 1.5X and 2.5X depending upon the amount of radio access spectrum available to the operator. The net result is a requirement to deploy a new backhaul technology that can deliver the necessary capacity, is packet based, and can easily add the new base stations as needed. Again, if fibre is present, it is the preferred technology. If fibre is not already present at the base station, however, the relative economics of fibre vs microwave must be taken into account.
Business Case for Fibre
The cost to deploy fibre is dominated by the installation expense and is thus distance sensitive: the longer the fibre lateral that must be constructed, the higher the cost of the backhaul. Microwave on the other hand does not significantly change in cost as distance increases, however there is an on-going annual charge for the tower space rental (if the towers are not owned by the operator) and the backhaul spectrum lease. The break-even distance for a fibre construction varies with local conditions, but is typically less than 1000 meters. Given the large number of new sites and the low fibre penetration, the majority of the base stations will be served by packet microwave, so it makes sense for us to look at packet microwave systems in more detail.
Business Case for Packet Microwave
The 10-year cost of ownership for packet microwave is only influenced by the capital cost of the radios to a minor extent (even though this tends to be the focus of the purchasing process). As is shown in the graphic above, the majority of costs are driven by lease costs for space and for spectrum. These costs are very dependant upon local regulatory conditions, and whether or not the operator owns the tower and site locations. A 10-year TCO analysis should therefore be done for every network that is considered. Current generation packet microwave systems have a number of features that can be used to mitigate these costs.
First of all, packet microwave systems are not limited to the SDH hierarchy of bit rates and can deliver throughput up to 50 Mbps per 7 MHz of spectrum with average sized packets (note that the throughput increases with smaller packets and some manufacturers quote these artificially high throughput rates. In practice the throughput with the average packet size is a much better measure of the real world system capability). Channel sizes are software defined and can be up to 56 MHz, if allowed by the regulator, for a throughput of 400 Mbps. Polarization multiplexing can be used to double this capacity, but at a cost per bit that is more than double. A feature known as adaptive modulation (the ability to adjust the modulation and/or coding to optimize the throughput under varying propagation conditions) allows these systems to deliver the maximum capacity under normal conditions and maintain the high priority traffic under poor conditions. Both of these translate to higher throughput, reduced antenna sizes and higher spectral utilization, resulting in a lower cost per bit.
The Importance of Ring and Mesh Network Topologies
Ring and mesh network topologies can further reduce the network cost per bit by decreasing the required redundancy costs, and minimizing the average antenna size. Traffic engineering, which allows the use of statistical multiplexing, makes use of all the available paths in a ring/mesh network, and leverages packet based prioritization to maintain priority traffic in the event of a failure condition. This can increase the effective network capacity by at least a factor of 4, further reducing the networks average cost per bit. Ease of installation and the reduction in site lease costs can be addressed by the use of all-outdoor system design, where the RF and base band electronics are integrated into a single outdoor unit, eliminating the need for co-location space in the cabinet. The net result of these factors is at least a 10-fold reduction of the network cost per bit.
The Future of Backhaul
Looking into the future is always subject to error, however we should expect a continuation of the trend towards smaller cell sizes in order to deliver higher capacity per user and make better use of the radio access spectrum. This will require ongoing innovation in the backhaul in terms of cost and integration levels. Capacity in excess of 1 Gbps per link is required in order to allow packet microwave to be used for the aggregation layer in the network and not just the final link to the end station. Traffic patterns that link base stations directly to one another, rather than hubbing all the traffic back to a central site (as proposed in the LTE standards) will further drive the need for ring/mesh network topologies rather than conventional hub and spoke designs. Finally spectral efficiency improvements at all channel sizes are required in order to deliver higher levels of network capacity without exhausting the available spectrum. We are by no means at the end of the road when it comes to innovation and evolution, if not outright revolution, of the backhaul network.
About the author: Alan Solheim is VP Product Management, DragonWave.
Share this article with others
Post to del.icio.us
Printed from http://www.eurocomms.com/features/113337/Backhaul_Evolution_-_The_elephant_in_the_room.html
Comment on this article
Skip to comments
We encourage users to analyse, comment on and even challenge European Communications's articles, including the one above - 'Backhaul Evolution - The elephant in the room'
User reviews and comments that include profanity or personal attacks or other inappropriate comments or material will be removed from the site.
Additionally, entries that are unsigned or contain "signatures" by someone other than the actual author will be removed. We will take steps to block users who violate any of our posting standards, terms of use or privacy policies or any other policies governing this site.