HetNets are hot. I hear about them in various contexts. Its difficult to find exactly what they are and how they will work though. There is a HetNets
special issue in IEEE Communications Magazine coming out next year but that's far away.
I found an interesting summary on HetNets in
Motorola Ezine that is reproduced below:
“The bigger the cell site, the less capacity per person you have,” said Peter Jarich, research director with market intelligence firm Current Analysis. “If you shrink coverage to a couple of blocks, you are having that capacity shared with a fewer number of people, resulting in higher capacity and faster data speeds.”
This is a topic the international standards body, the Third Generation Partnership Project (3GPP), has been focusing on to make small cells part of the overall cellular network architecture.
“What we’re seeing is a natural progression of how the industry is going to be addressing some of these capacity concerns,” said Joe Pedziwiatr, network systems architect with Motorola. “There is a need to address the next step of capacity and coverage by introducing and embracing the concepts of small cells and even looking at further advances such as better use of the spectrum itself.”
As such, discussion regarding this small-cell concept has emerged into what is called heterogeneous networks, or Het-Net, for short. The idea is to have a macro wireless network cooperating with intelligent pico cells deployed by operators to work together within the macro network and significantly improve coverage and augment overall network capacity. Small cells can also be leveraged to improve coverage and deliver capacity inside buildings. Indoor coverage has long been the bane of mobile operators. Some mobile operators are already leveraging this concept, augmenting their cellular service offering with WiFi access to their subscriber base in order to address the in-building coverage and capacity challenges facing today’s cellular solutions.
Pedziwiatr said this Het-Net structure goes far beyond what is envisioned for femtocells or standard pico cells for that matter. Introducing a pico cell into the macro network will address but just one aspect of network congestion, namely air interface congestion. The backhaul transport network may become the next bottleneck. Finally, if all this traffic hits the core network, the congestion will just have shifted from the edge to the core.
“This requires a system focus across all aspects of planning and engineering,” Pedziwiatr said. “We’re trying to say it goes beyond that of a femto. If someone shows up at an operator and presents a pico cell, that is just one percent of what would be needed to provide true capacity relief for the macro network.”
Femtocells, otherwise known as miniature take-home base stations, are obtained by end users and plugged into a home or office broadband connection to boost network signals inside buildings. A handful of 3G operators worldwide are selling femtocells as a network coverage play. For the LTE market, the Femtocell Forum is working to convince operators of the value of a femtocell when it comes to better signal penetration inside buildings and delivering high-bandwidth services without loading the mobile network. This is possible, because the backhaul traffic runs over the fixed line connection. However, this femtocell proposition largely relies on end user uptake of them—not necessarily where operators need them, unless they install femtocells themselves or give end users incentives to acquire them.
As with any new concept, there are challenges to overcome before Het-Nets can become reality. Het-Nets must come to market with a total cost of ownership that is competitive for an operator to realize the benefit of providing better capacity, higher data speeds, and most of all, a better end-user experience said Chevli.
“The level of total cost of ownership has to be reduced. That is where the challenge is for vendors to ensure that any new solution revalidates every existing tenet of cellular topology and evolve it to the new paradigm being proposed,” Chevli said. “You can’t increase the number of end nodes by 25X and expect to operate or manage this new network with legacy O&M paradigms and a legacy backhaul approach.”
One of the issues is dealing with interference and Het-Net network traffic policies. “How do you manage all of these small cell networks within the macrocell network?” asked Jarich. “Right now if you have a bunch of femtocells inside a house, there is this concept that the walls stop the macrocell signals from getting in and out. You get a separation between the two. Go outdoors with small cells underlying bigger cells and you get a lot more interference and hand-off issues because devices will switch back and forth based on where the stronger signal is.”
Pedziwiatr said for a Het-Net to work, it would require a change in node management, whereby an operator isn’t burdened with managing big clusters of small cells on an individual basis. “We see elements of SON (self organizing networks), self discovery and auto optimization that will have to be key ingredients in these networks. Otherwise operators can’t manage them and the business case will be a lot less attractive,” he said.
Fortunately, the industry has already been working with and implementing concepts of SON in LTE network solutions. In the femtocell arena also, vendors have been incorporating some elements and concepts of SON so that installing them is a plug-and-play action that automatically configures the device and avoids interference. But even then, Het-Nets will require further SON enhancement to deal with new use cases, such as overlay (macro deployment) to underlay (pico deployments) mobility optimization.
When it comes to LTE, SON features are built into the standard, and are designed to offer the dual benefit of reducing operating costs while optimizing performance. SONs will do this by automating many of the manual processes used when deploying a network to reduce costs and help operators commercialize their new services faster. SON will also automate many routine, repetitive tasks involved in day-to-day network operations and management such as neighbor list discovery and management.
Other key sticking points are deployment and backhaul costs. If operators are to deploy many small cells in a given area, deploying them and backhauling their traffic should not become monumental tasks.
Chevli and Pedziwiatr envision Het-Nets being deployed initially in hot zone areas – where data traffic is the highest – using street-level plug-and-play nodes that can be easily installed by people with little technical know-how.
“Today, macro site selection, engineering, propagation analysis, rollout and optimization are long and expensive processes, which must change so that installers keep inventories of these units in their trucks, making rollout simple installations and power-ups,” said Pedziwiatr. “These will be maintained at a minimum with quick optimization.”
The notion of backhauling traffic coming from a large cluster of Het-Net nodes could also stymie Het-Nets altogether. Chevli said that in order to keep costs down, Het-Net backhaul needs to be a mix of cost-effective wireless or wired backhaul technology to aggregate traffic from what likely will be nodes sitting on lamp posts, walls, in-building and other similar structures. The goal then is to find a backhaul point of presence to aggregate the traffic and then put that traffic on an open transport network in the area.
Backhaul cost reductions may also be a matter of finding ways to reduce the amount of backhaul forwarded to the core network, Pedziwiatr said. These types of solutions are already being developed in the 3G world to cope with the massive data traffic that is beginning to crush networks. For traffic such as Internet traffic, which doesn’t need to travel through an operator’s core network, offloading that traffic as close to the source as possible would further drive down the cost of operation through the reduction of backhaul and capacity needs of the core network.
In the end, with operators incorporating smaller cells as an underlay to their macro network layer rather than relying on data offloading techniques such as femtocells and WiFi that largely depend on the actions of subscribers and impacted by the surrounding cell operating in the same unlicensed frequency, Het-Nets in licensed spectrum will soon become the keystone in attacking the ever-present congestion issue that widely plagues big cities and this is only likely to get worse over time.
Just went through Alcatel-Lucent strategic paper on whether to go for more Macrocell sites or rather have Metrocells instead. 
There are some interesting blog posts on Broadband Traffic Managemenet on Backhaul. Here are few excerpts:
The first one, 'Guidelines for LTE Backhaul Traffic Estimation' document describes how a model is developed to predict traffic levels in transport networks used to backhaul LTE eNodeBs. Backhaul traffic is made up of a number of different components of which user plane data is the largest, comprising around 80-90% of overall traffic, slightly less when IPsec encryption is added. These results reveal that the cell throughput characteristics for data carrying networks are quite different to those of voice carrying networks.
The purpose of second one, 'NGMN Whitepaper LTE Backhauling Deployment Scenarios' is to support operators in their migration from current architectures to new, packet-based backhaul networks. With the introduction of LTE operators need to look at how the backhauling network, the network domain that connects evolved NodeBs (eNBs) to MME and S/P-GW, is capable of adapting to the new requirements, namely the adoption of a packet infrastructure, without disrupting the existing services. This paper introduces some reference architectures, moving from a pure layer 2 topology to a full layer 3 one, discussing some elements to be considered in the design process of a network.
