Meanwhile, the startup companies that make the gear over which those services traverse have seen investment fall from $796 million in 2009 annually to just $270 million in the 12 months ending in June, Ovum found. According to Ovum principal analyst Matt Walker:

“A funding disconnect has thereby emerged between network builders and network users. Lots of innovation and venture capital is targeting the network users, such as mobile apps and OTT platforms. However, little of it is directly helping the network builders. With a weak start-up pipeline, the industry relies more on incumbent vendors to generate new ideas and products. Their budgets are bigger, but VCs are often better at funding ‘game changing’ ideas ignored by established vendors.”

Admittedly, investing in the next big social network or an app that could generate millions of downloads is a lot sexier than, say, envelope tracking technology or cell site radio frequency filters. But those infrastructure innovations are just as important. The capabilities of many apps and services have already far exceeded the ability of our mobile networks to deliver those apps and services at a reasonable cost (think Netflix on 4G tablet). If we let network innovation slip, we could wind up with a bunch of very powerful services that have nowhere to go.

As Walker points out, the onus for innovation thus falls on the big established telecom vendors, and it’s quite the burden. Ovum estimates that with the falloff in startup investment, big network infrastructure makers’ R&D budgets are now 90 times larger than the investment going into networking startups –- that’s up from 30X two years ago.

Nokia Siemens Networks’ conception of a heterogeneous network

Don’t get me wrong — the Ciscos, Ericssons and Huaweis of the world are responsible for some amazing science and innovation. And today they’re building the small cell and heterogeneous networks of the future. But there are limits to what the big vendors can accomplish. The R&D budgets of the big industrial labs have shrunk immensely in the last two decades, and there’s only so much talent and so many resources those vendors can devote to innovation.  The biggest issue, though, is that the big equipment makers innovate in much different ways than small startups.

Big vendors have big ingrained investments

Look around. A lot of the wired and wireline networks we use on a daily basis have been with us for a while. The first 2G networks in the US went up in the late 1990s and they’re largely still in use. A good part of the big vendors’ businesses is maintaining, upgrading and iterating on the networks they’ve already built.

The lightRadio Cube, Alcatel-Lucent’s vision for the small cell.

That doesn’t mean the big vendors are merely redesigning the same old equipment, but they’re definitely looking for continuity with their older networks. Alcatel-Lucent’s lightRadio and Nokia Siemens’ Liquid Radio architectures, for instance, are truly mind-blowing approaches to the new heterogeneous network, but they’re still fundamentally the cellular technologies that have been these vendors’ bread and butter since the birth of wireless.

When Wi-Fi came along as a mobile data alternative to cellular, these vendors were resistant if not outright hostile. It took two startups, BelAir Networks and Ruckus Wireless to make the business case to carriers for large-scale outdoor Wi-Fi networks to supplement 3G and 4G networks.

The big vendors are working largely within global standards frameworks. That’s by no means a bad thing. It’s why an iPhone can communicate with a Nokia-built base station, and a Cisco router can be plugged into an Ericsson core network. But standards work is painfully slow. A lot of the innovation work in networking technology works goes on outside of the standards bodies, and if that work proves successful it wind up shaping the standards themselves.

There’s probably no better example in wireless than CDMA. Qualcomm’s upstart cellular interface was initially adopted by a single US carrier, AirTouch, but it eventually became the basis for all global 3G networks.

Innovating between the lines

While the big vendors have focused on the overarching evolution of networks it’s up to infrastructure core technology startups to fill in technology gaps. Companies like NSN and Ericsson will most certainly handle the large-scale rollout of small cells and hetnets in the future, just like Apple and Samsung will be designing our future 4G smartphones and connected tablets.

But it will be startups like Seattle’s still under-the-radar PivotBeam that are developing the critical software defined antennas that will link these millions of small cells back to the network core. And it will be small engineering companies like Nujira and Quantance supplying the power envelope tracking technology giving those 4G phones a tolerable battery life.

I’m not saying all of these specific companies are all going to be the next Qualcomm, and that you should go invest in them. But they’re part of a critical network infrastructure startup scene, and that scene appears to be shrinking. We’re already starting to see the consequences. The industry has started delivering speed in the form of LTE but it has so far failed to deliver us the cheap capacity critical to moving the mobile industry forward. If the investors keep neglecting network startups, that problem is only going to get worse.

Photo courtesy of Shutterstock user Mati Nitibhon

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• Mobile 2012 and beyond
• Who and what to watch in the new era of the living room
• New solutions for the evolving mobile network

Big LTE wins in Japan and Korea have turned Asia into NSN’s biggest market region, surpassing Europe and accounting for €1.27 billion of its €3.5 billion in sales last quarter. What’s more, some of these Asian deals make use of some of LTE’s most sophisticated capabilities. SK Telecom is building an LTE-Advanced network that incorporates new small cell and self-optimizing network (SON) technologies unseen on any other 4G network globally.

NSN has tried to cultivate that burgeoning reputation of being on the cutting edge of 4G. In recent weeks it’s launched a slew of new services and products, all designed to transform the network from today’s grids of spaced-out towers into future heterogeneous networks. These HetNets will combine big and small cells, mix and match different radio technologies ranging from cellular to Wi-Fi and provide a dense layer of capacity under the macro network’s coverage umbrella.

Small networks mean big capacity gains

Nokia Siemens Networks’ conception of a heterogeneous network

These technologies, which NSN calls Flexi Zone and Liquid Net, may sound like strange engineering concepts, but Nokia Siemens is trying to define them in ways that has meaning to the average consumer. It’s set a goal of creating a network architecture that will increase the overall capacity of our mobile networks by 1,000 times. Simultaneously, the cost of delivering that capacity will drop to the point that they typical consumer can increase their data usage by a factor of 10 for the same costs they pay today, said Bill Payne, NSN CTO for North America.

“We’re saying by 2020, we’ll have 1 GB per day per user for under $1 in cost,” Payne said. That amounts to 30 GB a month for $30, which is the price a typical U.S. consumer pays for a 2 to 3 GB data plan (though they actually consumer less than 1 GB each month).

Talking about 1,000x increase in network capacity is nothing new. NSN’s archrival Ericsson has been talking up the concept for years and Qualcomm recently detailed the chip vendor’s own “1000x challenge” in a GigaOM post. But NSN is laying out a detailed path on just how the industry gets there. According to Payne, the industry will achieve that kind of capacity in three stages:

• A 10x increase from better cellular technologies. “The first order will be spectral efficiency,” Payne said. “It’s the logical starting point because that’s the work people in this industry has[HAVE?] been doing most of their professional lives.” CDMA and HSPA networks are giving way to LTE networks, which will themselves give way to LTE-Advanced networks. There are limits to how the industry can innovate in this direction though – only so much data that can be shoved into a hertz of spectrum, Payne said.
• A 10x increase from new spectrum. While this aspect is out of NSN’s control, carriers worldwide are working with regulators to devote new airwaves for mobile broadband. AT&T just cleared its 2.3 GHz licenses for LTE use, and the FCC is working to clear more broadcast spectrum for cellular use. Dumping more frequencies into our networks may not be the most elegant solution, but it’s definitely the direction most carriers would rather grow in.
• A 10x increase from small cells. Changing the fundamental topologies of our networks will allow carriers to make that final critical leap to 1,000x, Payne said. The principle is a simple one. By putting 10 cells in the space occupied by a single big one you get close to a 10x increase in capacity. By layering in Wi-Fi and other radio technologies you get even bigger gains. And by creating a network in which a single device can link to multiple radio nodes, the capacity consistently available to individual users increases dramatically.

Building on NSN’s LTE gains

NSN is pushing its LTE-Advanced and HetNet concepts hard, promising its customers that it can not only build their 4G networks today but turn them into the 4G networks of tomorrow. In Arlington Heights, Ill., it’s built a live small cell test network using public safety spectrum to test out the HetNet’s performance in real-world environments.

Its efforts seem to be paying off. Though Ericsson and Huawei still dominate the global telecom infrastructure market, ABI Research recently ranked NSN the world’s leading vendor in LTE contracts, intellectual property and, most notably, its progress in small cells.

LTE image courtesy of Shutterstock user Inq

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• New solutions for the evolving mobile network
• Mobile 2012 and beyond
• CES 2012: a recap and analysis

Sounds like science fiction, but this is real technology commercially available today. Israeli startup Intucell has already deployed it in its home country with operator Pelephone and is engaged in multiple other trials. AT&T, however, is the big fish, and Intucell is hoping its nationwide launch of the technology across its HSPA and LTE networks will validate SON for the rest of the world.

I detailed how Intucell’s dynamic SON platform works in December, but in short, it uses a distributed network intelligence to track the network’s health and levels of congestion. It then adjusts the transmission power of each cell in the network to create the best possible configuration for both coverage and capacity. Quite literally cell towers start following you, expanding their cell radii as you move closer to their edges, while neighboring cells recede. By moving the network around you as you yourself move through the network, SON can find the optimal overall topology at any given movement to provide the best coverage and capacity to thousands of users within a cluster of cells. It’s pretty cool stuff.

So what’s the benefit? According to AT&T, its initial trials of the technology over HSPA networks in two markets resulted in a 10 percent reduction in dropped calls. That is a tangible benefit, but it is nothing when you compare it to SON’s impact on mobile data capacity. AT&T isn’t releasing any numbers on how SON improved data performance, but Intucell pointed to data it has collected from other trials.

According to Intucell CEO Rani Wellingstein, the technology can reduce cell congestion anywhere from 10 to 40 percent depending on the configuration of the network, allowing operators to pack more capacity onto less infrastructure. Those capacity increases are also passed onto the consumer in the form of faster speeds. SON can boost the average throughput by up to 15 percent of the network’s theoretical limit, which in many cases could equate to a 1 Mbps or greater increase in downlink speeds.

“AT&T is perceived as the operator with the highest data crunch problem,” Wellingstein said. “It has the highest penetration of smartphones and the highest penetration of iPhones. It makes sense that it would be the first operator in North America to deploy dynamic SON.”

SON has other benefits. It can make networks self-healing. If a cell site is down — which, according to Wellingstein, is the case for 1 to 2 percent of the world’s cells at any given time — the surrounding cells can expand their radius to fill the hole in the network. When the site is repaired the cells retract to their normal size.

AT&T is moving aggressively with the technology. It started trialing Intucell’s platform last April, but it plans to have the technology in all of its networks nationwide by the end of the year. Intucell and AT&T did not disclose the financial terms of their deal, but Wellingstein said deploying SON is a fraction of the cost of adding the same capacity through new radio infrastructure. He said a U.S. operator could add basic SON functions to a nationwide network for around $50 million. Not exactly cheap, but in a country the size of the U.S., even the most modest network deployment can run multiple billions of dollars.

Related research and analysis from GigaOM Pro:
Subscriber content. Sign up for a free trial.

• New solutions for the evolving mobile network
• Mobile 2012 and beyond
• LTE-Advanced: what it is and isn’t